MAFS 5030 — Quantitative Modeling of Derivative Securities

Topic 1 – Introduction to Derivative Instruments

1.1 Basic instruments: bonds, forward contracts and futures

1.2 Exotic swap products

1.3 Options: rational boundaries of option values

1.4 American options: Optimal early exercise policies

1
1.1 Basic derivative instruments: bonds, forward contracts,
swaps and options

A bond is a debt instrument requiring the issuer to repay to the

lender/investor the amount borrowed (par or face value) plus inter-
ests over a specified period of time.

Specify (i) the maturity date when the principal is repaid;

(ii) the coupon payments over the life of the bond
P

stream of coupon payments

maturity
2
 • The coupon rate offered by the bond issuer represents the cost
of raising capital. It depends on the prevailing risk free interest
rate and the creditworthiness of the bond issuer. It is also af-
fected by the values of the embedded options in the bond, like
the conversion right in a convertible bond.

• Assume that the bond issuer does not default or redeem the
bond prior to maturity date, an investor holding this bond until
maturity is assured of a known cash flow stream. This explains
why bond products are also called the Fixed Income Product-
s/Derivatives.

Pricing of a bond

Based on the current information of the term structure of interest

rates (yield curve) and the embedded option provisions, find the fair
price that the bond investor should pay at the current time so that
the deal is fair to both counterparties. Also, potential default losses
(probability of default, exposure and recovery value upon default)
should be taken into account in computing the fair bond price.
3
Features in bond indenture

1. Floating rate bond

The coupon rates are reset periodically according to some pre-
determined financial benchmark, like LIBOR + spread, where
LIBOR is the LONDON INTER-BANK OFFERED RATE.

2. Amortization feature – principal repaid over the life of the bond.

3. Callable feature (callable bonds)

The issuer has the right to buy back the bond at a specified
price. Usually this call price falls with time, and often there
is an initial call protection period wherein the bond cannot be
called.

4
4. Put provision – grants the bondholder the right to sell back to
the issuer at par value on designated dates.

5. Convertible bond – gives the bondholder the right to exchange

the bond for a specified number of shares of the issuer’s firm.

• Bondholders can take advantage of the future growth of the

issuer’s company.
• Issuer can raise capital at a lower cost.

6. Exchangeable bond – allows the bondholder to exchange the

bond par for a specified number of common stocks of another
corporation.

5
Short rate

Let r(t) denote the short rate, which is in general stochastic. This
is the interest rate that is applied over the next infinitesimal ∆t time
interval (t, t + ∆t]. The short rate is a mathematical construction,
not a market reality.

Money market account: M (t)

An investor puts $1 at time t and let it earn interest at the rate

r(t) continuously over the period (t, T ). The governing differential
equation of M (t) is given by

dM (t) = r(t)M (t) dt.

Here, dM (t) represents the interest accrued over (t, t + dt).

6
 ∫ T ∫ T
dM (u)
= r(u) du
t M (u) t
so that
∫T
r(u) du
M (T ) = M (t)e t .
∫T
Here, e t r(u) du is seen to be the growth factor of the money mar-
ket account over [t, T ] based on continuous compounding. If r is
constant, then

growth factor = erτ , τ = T − t.

If r(t) is stochastic, then M (T ) is also stochastic.
∫
− tT r(u) du
The reciprocal of the growth factor e is called the discount
factor.

7
Discount bond price: B(t, T )

τ = T − t = time to bond’s maturity

The price that an investor on the zero-coupon (discount) bond with

unit par is willing to pay at time t if the bond promises to pay him
back $1 at a later time (maturity date) T .

This fair value is called the discount bond price B(t, T ), which is
given by the expectation
[ ∫
of the ]discount factor based on the cur-
− tT r(u) du
rent information: Et e . If r is constant, then B(t, T ) =
e−rτ , τ = T − t.

8
Forward contract

The buyer of the forward contract agrees to pay the delivery price
K dollars at future time T to purchase a commodity whose value at
time T is ST . The pricing question is how to set K?

How about
E[exp(−rT )(ST − K)] = 0
so that K = E[ST ]?

This is the expectation pricing approach, which cannot enforce a

price. When the expectation calculation E[ST ] is performed, the
distribution of the asset price process comes into play.

9
Objective of the buyer:

To hedge against the price fluctuation of the underlying commodity.

• Intension of a purchase to be decided earlier, actual transaction

to be done later.

• The forward contract needs to specify the delivery price, amoun-

t, quality, delivery date, means of delivery, etc.

Potential default of either party (counterparty risk): writer or buy-

er.

Forward contract is the simplest form of a financial derivative. The

exposure to the writer (short position) of the forward can be hedged
by long holding one unit of the underlying asset. In this case, the
hedge ratio is one.

10
Terminal payoff from a forward contract

K = delivery price, ST = asset price at maturity

This is a zero-sum game between the writer (short position) and

buyer (long position). For cash settlement of the terminal payoff,
one can sell the underlying asset immediately to realize the cash
value ST upon maturity of the contract.
11
Replication enforces forward price

Is the forward price related to the expected price of the commodity

on the delivery date? Provided that the underlying asset can be
held for hedging by the writer, then

forward price
= spot price + cost
| of fund +
{z storage cost}
cost of carry

• Upfront cost (through borrowing) is required to acquire the un-

derlying commodity at the spot price. Cost of fund is the interest
costs accrued over the period of the forward contract.

• Cost of carry is the total cost incurred to acquire and hold the
underlying asset, say, including the cost of fund and storage
cost.

• Dividends paid to the holder of the asset are treated as negative

contribution to the cost of carry.

12
Numerical example on arbitrage

– spot price of oil is US$19

– quoted 1-year forward price of oil is US$25
– 1-year US dollar interest rate is 5% pa
– storage cost of oil is 2% per annum, paid at maturity

Any arbitrage opportunity? Yes

Sell the forward and expect to receive US$25 one year later. Borrow
$19 now to acquire oil, pay back $19(1+0.05) = $19.95 a year later.
Also, one needs to spend $0.38 = $19 × 2% as the storage cost.

total cost of replication (dollar value at maturity)

= spot price + cost of fund + storage cost
= $20.33 < $25 to be received.
Close out all positions by delivering the oil to honor the forward. At
maturity of the forward contract, guaranteed riskless profit = $4.67.
13
How to perform replication when the underlying is a stock index?

• Long a basket of stocks whose composition is the same as the

weights that calculate the stock index.

• Upon maturity of the index futures (traded in an exchange), the

cash settlement is realized by the liquidation of the basket of
stocks.

However, since the settlement futures value is the average of the

stock index taken at 5-minute time intervals on the settlement date
of the index futures, liquidation has to be executed at every 5-
minute time interval accordingly. There may be a potential slippage
since the actual implementation of liquidation may take 30 seconds
or more. Also, there are uncertainties with regard to the dividends
received from the basket of stocks.

14
Value and price of a forward contract

Let f (S, τ ) = value of forward, F (S, τ ) = forward price,

τ = time to expiration,
S = spot price of the underlying asset.
Further, we let

B(τ ) = value of an unit par discount bond with time to maturity τ

• If the interest rate r is constant and interests are compounded

continuously, then B(τ ) = e−rτ .

• Assuming no dividend to be paid by the underlying asset and no

storage cost.

We construct a “static” (fix the hedge ratio to be one throughout

the life of the forward contract) replication of the forward contract
by a portfolio of the underlying asset and bond.
15
Portfolio A: long one forward and a discount bond with par value K

Portfolio B: one unit of the underlying asset

Both portfolios become one unit of asset at maturity. Let ΠA(t)

denote the value of Portfolio A at time t. Note that ΠA(T ) = ΠB (T ).
By the “law of one price”,∗ we must have ΠA(t) = ΠB (t). The
forward value is given by

f = S − KB(τ ).

The forward price is defined to be the delivery price which makes

f = 0, so K = S/B(τ ). Hence, the forward price is given by

F (S, τ ) = S/B(τ ) = spot price + cost of fund.

∗ Suppose ΠA (t) > ΠB (t), then an arbitrage can be taken by selling Portfolio A
and buying Portfolio B. An upfront positive cash flow is resulted at time t but
the portfolio values are offset at maturity T . Failure of law of one price implies
the existence of an arbitrage opportunity.
16
Failure of the law of one price leads to existence of an arbitrage op-
portunity (an important financial economic concept to be discussed
in Topic 2)

ODZRIRQHSULFH

DEVHQFHRI
DUELWUDJH

Can you find an example where law of one price is observed while
there exists an arbitrage opportunity?

17
Forward price formula with discrete carrying costs

Suppose an asset has a holding cost of c(k) per unit in period k,

k = 1, 2, . . . , M , and the asset can be sold short. Suppose the initial
spot price is S. The theoretical forward price F is
M
∑
S c(k)
F = + ,
B(0, M ) E [d(k, M )]
k=1 0
where the market expectation of the discount factor d(k, M ) at time
zero is given by

E0[d(k, M )] = B(0, M )/B(0, k).

Here, B(0, k) denotes the market observable time-0 price of a dis-
count bond that matures at time k.

The terms on the right hand side represent the future value at
maturity of the total costs required for holding the underlying asset
for hedging.

18
Discrete dividend paying asset

We may treat the carrying charges c(k) as negative dividends D(k),

k = 1, 2, . . . , M . Let D be present value of all dividends received
from holding the asset during the life of the forward contract, where
M
∑
D= B(0, k)D(k), then
k=1
M
∑
S D(k)B(0, k) S−D
F = − = .
B(0, M ) k=1 B(0, M ) B(0, M )

19
Alternative proof

We modify Portfolio B to contain one unit of the asset plus bor-

rowing of D dollars. The loan of D dollars will be repaid by the
dividends received by holding the asset. We then have

f + KB(τ ) = S − D
so that
f = S − [D + KB(τ )].
Setting f = 0 to solve for K, we obtain F = (S − D)/B(τ ).

The “net” asset value is reduced by the amount D due to the

anticipation of the dividends. Unlike holding the asset, the holder
of the forward will not receive the dividends. As a fair deal, he
should pay a lower delivery price at forward’s maturity.

20
Cost of carry

Additional costs to hold the commodities, like storage, insurance,

deterioration, etc. These can be considered as negative dividends.
Treating U as −D, we obtain

F = (S + U )erτ ,
U = present value of total cost incurred during the remaining life of
the forward to hold the asset.

Suppose the costs are proportional and paid continuously, we have

F = Se(r+u)τ ,
where u = cost per annum as a proportion of the spot price.

In general, F = Sebτ , where b is the cost of carry per annum. Let q

denote the continuous dividend yield per annum paid by the asset.
With both continuous holding cost and dividend yield, the cost of
carry b = r + u − q.

21
Proportional carrying charge

Forward contract is written at time 0 and there are M periods until

delivery. The carrying charge in period k for one unit of stock is
qS(k) to be paid at time k, where q is a proportional constant. Show
that the forward price is
S(0)/(1 − q)M
F = .
d(0, M )

• The above forward price formula becomes F = Se(r+u)τ when we

change from discrete compounding to continuous compounding,
1
where d(0,M and 1 rτ and euτ , respectively.
) M are related to e
(1−q)

• We expect the forward price increases when the carrying charges

become higher since the cost of setting up the static replicating
portfolio is higher.

22
Borrow αS(0) dollars at current time to buy α units of assets and
long one forward. Sell out q portion of asset at each period in order
to pay for the carrying charge. After M period, α units becomes
α(1 − q)M .

The goal is to make available one unit of the asset for delivery at
1
maturity. We set α(1 − q)M to be one and obtain α = .
(1 − q) M

The portfolio of longing the forward with delivery price K and a

bond with par K is equivalent to long α units of the asset. This
gives
f + Kd(0, M ) = αS(0).
By setting f = 0 to obtain the forward price K, we obtain
/
S(0)
K= d(0, M ).
(1 − q)M

23
Futures contracts න೛Ӕँ

A futures contract is a legal agreement between a buyer (seller) and

an established exchange or its clearing house in which the buyer
(seller) agrees to take (make) delivery of a financial entity at a
specified price at the end of a designated period of time. Usually
the exchange specifies certain standardized features.

Mark to market the account

Pay or receive from the writer the change in the futures price
through the margin account so that payment required on the ma-
turity date is simply the spot price on that date.

Credit risk is limited to one-day performance period

24
Roles of the clearing house and margin account

• Minimize the counterparty risk through the margin account.

• Provide the platform for parties of a futures contract to unwind

their position prior to the settlement date.

Margin requirements

Initial margin – paid at inception as a deposit for the contract.

Maintenance margin – minimum level before the investor is required

to deposit additional margin.

25
Example (Margin)

Suppose that Mr. Chan takes a long position of one contract in

corn (5, 000 kilograms) for March delivery at a price of $2.10 (per
kilogram). And suppose the broker requires margin of $800 with a
maintenance margin of $600.

• The next day the price of this contract drops to $2.07. This
represents a loss of 0.03 × 5, 000 = $150. The broker will take
this amount from the margin account, leaving a balance of $650.
The following day the price drops again to $2.05. This repre-
sents an additional loss of $100, which is again deducted from
the margin account. As this point the margin account is $550,
which is below the maintenance level.

• The broker calls Mr. Chan and tells him that he must deposit at
least $250 in his margin account, or his position will be closed
out.

26
Dynamic strategy that replicates the daily margin settlement
in a futures contract

Consider an asset with price S̃T at time T . An investor who pays an

amount Gt,T that equals the futures price of the asset and together
performs the dynamic strategy of long holding futures on successive
dates is equivalent to pay the time-t spot price of a security which
has a payoff
SeT
e e
Bt,t+1B t+1,t+2 · · · BT −1,T
at time T . Note that quantities with “tilde” at top indicate stochas-
tic variables. Note that the payoff is seen to be some units of the un-
1
derlying asset, where the number of units is e e
Bt,t+1B t+1,t+2 · · · BT −1,T
(this is the same as the value of the money market account starting
at $1 and accumulating over the period [t, T ]).

27
The dynamic strategy is presented as follows.

• We start with B 1 long futures contracts at time t. On day

t,t+1
τ , the investor long holds 1 futures contracts. The
e
Bt,t+1 ···Bτ,τ +1
gain/loss from the futures position on day τ earns/pays the
overnight rate e 1 .
Bτ,τ +1

• Also, invest the dollar amount of Gt,T in a one-day risk free bond
and roll the cash position over on each day at the one-day rate.
This is like “rolling over” in a money market account on a daily
basis.

• The investment of Gt,T is equivalent to the price paid to acquire

the asset.

28
As an illustrative example, take t = 0 and T = 3.

1. Take 1/B0,1 long futures at t = 0;

1/B0,1B1,2 long futures at τ = 1;
1/B0,1B1,2B2,3 long futures at τ = 2.
Note that B1,2 and B2,3 are market observable bond prices at
τ = 1 and τ = 2, respectively. Also, it causes nothing to change
the holding of units of futures on successive dates.

2. Invest the dollar amount of G0,3 in one-day risk free bond and
roll over the net cash position

Time Profits from futures Bond position Net position

0 — G0,3 G0,3
G0,3 G1,3
1 1
B0,1
(G1,3 − G0,3 ) B0,1 B0,1
G1,3 G2,3
2 1
B0,1 B1,2
(G2,3 − G1,3 ) B0,1 B1,2 B0,1 B1,2
G2,3 G3,3
3 1
B0,1 B1,2 B2,3
(G3,3 − G2,3 ) B0,1 B1,2 B2,3 B0,1 B1,2 B2,3
= S3
B0,1 B1,2 B2,3

Note that G3,3 = S3.

29
The forward price of the asset, Ft,T , is the time-t spot price of an
asset which has a payoff SeT /Bt,T at time T .

Proof

Change the daily settlement in a futures to have one single settle-

ment at T in a forward. This is done by long holding 1/Bt,T units
of forward at time t and invest Ft,T on a discount bond at time t.
The portfolio ends up to become SeT /Bt,T at time T .

Remark
Since Bt,T and St are known at time t, Ft,T is the time-t spot price
1
of a security that pays units of the underlying asset. Hence,
Bt,T
St
we have Ft,T = .
Bt,T

30
Pricing issues

We consider the discrete-time model and assume the existence of

a risk neutral pricing measure Q. We have demonstrated the con-
struction of the dynamic replication strategy. The fair time-t price
SeT
of e
is the cost of setting up the replicating port-
Bt,t+1 · · · BT −1,T
folio, which is Gt,T . On the other hand, based on the risk neutral
valuation principle, the time-t price of a security is given by the
Q-expectation of the discounted terminal payoff at T , where
 
e e SeT
Gt,T = EQ Bt,t+1B t+1,t+2 . . . BT −1,T e e e

Bt,t+1B t,t+1 Bt+1,t+2 · · · BT −1,T
= EQ[SeT ].
The result remains valid for the continuous-time counterpart. Under
the continuous-time model, the risk neutral valuation principle gives
the time-t price of a contingent claim as follows:
[ ∫T ]
Vt = EQ e− t r(u) du
VT .

31
Difference in futures price Gt and forward price Ft

Difference in payment schedules may lead to difference in futures

and forward prices since different interest rates are applied on in-
termediate payments. When the interest rates are deterministic,
we have Gt,T = Ft,T . This is a sufficient condition for equality of
the two prices. The necessary and sufficient condition is that the
discount process and the underlying price process are uncorrelated
under the risk neutral measure Q.

• When physical holding of the underlying index (say, snow fall

amount in a ski resort) for hedging is infeasible, then the buyer
sets
forward price = EP [ST ],
where P is the subjective probability measure of the buyer.

32
• When the physical holding of the asset is subject to daily set-
tlement through the margin requirement (dynamic rebalancing)

futures price = EQ[ST ],

where Q is the risk neutral measure that uses the money market
account as the numeraire. As a remark, 1 is
Bt,t+1 B̃t+1,t+2 ···B̃T −1,T
seen as the money market account rolling daily over the period
S̃T
[t, T ]. The asset when normalized by the
Bt,t+1 B̃t+1,t+2 ···B̃T −1,T
money market account becomes S̃T .

33
By the risk neutral valuation principle, we have
[ ∫T ] [ ∫T ]
B(t, T ) = EQ e− t r(u)du
and St = EQ e− t r(u) du
ST

so that
St
Gt − Ft = EQ [ST ] −
B(t, T )
[ ∫T ] [ ∫T ]
EQ[ST ]EQ e− t r(u) du
− EQ e− t r(u) du
ST
=
B(t, T )
[ ∫T ]
covQ e− t r(u) du
, ST
= − .
B(t, T )

The spread between Gt and Ft becomes zero when the discount pro-
cess and the price process of the underlying asset are uncorrelated
under the risk neutral measure Q. In the special case where interest
rates are deterministic, we have equality of Gt and Ft.

34
Currency forward

The underlying is the exchange rate X, which is the domestic cur-

rency price of one unit of foreign currency.

rd = constant domestic interest rate

rf = constant foreign interest rate

Portfolio A: Hold one currency forward with delivery price K

and a domestic bond of par K maturing on the
delivery date of forward.

Portfolio B: Hold a foreign bond of unit par maturing on the

delivery date of forward.

Holding of the domestic and foreign bonds allow the bonds to earn
the interest rate in the respective currency.

35
Remark

The underlying asset of a foreign currency forward is one unit of

foreign currency (in the form of market account) that pays dividend
yield rf .

Let ΠA(t) and ΠA(T ) denote the value of Portfolio A at time t and
T , respectively. On the delivery date, the holder of the currency
forward has to pay K domestic dollars to buy one unit of foreign
currency. Hence, ΠA(T ) = ΠB (T ), where T is the delivery date.

Using the law of one price, ΠA(t) = ΠB (t) must be observed at the
current time t.

36
Interest rate parity relation

Note that
Bd(τ ) = e−rdτ , Bf (τ ) = e−rf τ ,
where τ = T − t is the time to expiry. Let f be the value of the
currency forward in domestic currency,

f + KBd(τ ) = XBf (τ ),
where XBf (τ ) is the value of the foreign bond in domestic currency.

By setting f = 0, the forward price of the currency forward is

XBf (τ )
K= = Xe(rd−rf )τ .
Bd(τ )
We may recognize rd as the cost of fund and rf as the dividend
yield. This result is the well known Interest Rate Parity Relation.

37
Flexible notional currency forward

Consider a 6-month forward contract. The exchange rate over each

one-month period is preset to assume some constant value.

The holder can exercise parts of the notional at any time during the
life of the forward, but she has to exercise the whole notional by
the maturity date of the currency forward.

Normally, we set Fj = F1e(rd−rf )(tj −t1), j = 2, 3, . . . , 6. How to set

F1 so that it is fair to both parties? We expect F1 > Xe(rd−rf )t1
since the buyer needs to pay more domestic currency to buy one
unit of foreign currency. The extra amount represents the premium
of this flexibility.

38
As the optimal policy is independent of the notional, the holder
chooses either no action or exercise the full notional (partial exercise
is non-optimal). This is called the bang-bang strategy.

If otherwise, for any optimal proportion of exercise, α < 1, the

remaining 1 − α is also subject to an optimal exercise of α portion,
and so forth. It ends up with the exercise of the full notional.

Remark
Suppose differential transaction costs are charged according to the
notional amount, say 0.1% on the first $10, 000 and 0.08% on the
next $10, 000,etc., then the property of independence of the notional
fails. However, if the transaction cost is proportional to the notional,
then the independence property remains.

As part of the pricing issue, one has to determine the threshold ex-
change rate above which it is optimal to exercise the whole notional.

39
Bond forward

The underlying asset is a zero-coupon bond of maturity T2 with the

settlement date T1, where t < T1 < T2.

The holder pays the delivery price F of the bond forward on the
forward maturity date T1 to receive a bond with par value P on the
maturity date T2.

40
Bond forward price in terms of traded bond prices

Let Bt(T ) denote the traded price of unit par discount bond at
current time t with maturity date T .

Present value of the net cashflows

= −F Bt(T1) + P Bt(T2).
To determine the forward price F , we set the above value zero and
obtain
F = P Bt(T2)/Bt(T1).
Here, P Bt(T2) can be visualized as the spot price of the discount
bond. The forward price is given in terms of the known market
bond prices observed at time t with maturity dates T1 and T2.

41
Forward on a coupon-paying bond

The underlying is a coupon-paying bond with maturity date TB .

Note that the bond is a traded security whose value changes with
respect to time.

Let TF be the delivery date of the bond forward, where TF < TB . Let
ti be the coupon payment date of the bond on which deterministic
coupon ci is paid. Let t be the current time, where t < TF < TB .
Some of the coupons have been paid at earlier times. Let F be the
forward price, the amount paid by the forward contract holder at
time TF to buy the bond.

42
At TB , the bondholder receives par plus the last coupon.

43
Based on the forward price formula: F = S−D
B(τ )
, we deduce that

spot price of bond c4Bt(t4) c5Bt(t5)

F = − − .
Bt(TF ) Bt(TF ) Bt(TF )

Let P be the par value of the bond. After receiving the bond at
TF , the bond forward holder is entitled to receive c6, c7 and P once
he has received the underlying bond. By considering the cash flows
after TF , he pays F at TF and receives c6 at t6, c7 + P at TB .

Present value of cash flows at time t

= −F Bt(TF ) + c6Bt(t6) + c7Bt(TB ) + P Bt(TB ).
Hence, the bond forward price is given by
c Bt(t6) + c7Bt(TB ) + P Bt(TB )
F = 6 .
Bt(TF )
Remark
Equating the two expressions gives the spot price of the bond in
terms of the cash flows.

44
Example — Bond forward

• A 10-year bond is currently selling for $920.

• Currently, hold a forward contract on this bond that has a de-

livery date in 1 year and a delivery price of $940.

• The bond pays coupons of $80 every 6 months, with one due
6 months from now and another just before maturity of the
forward.

• The current interest rates for 6 months and 1 year (compound-

ed semi-annually) are 7% and 8%, respectively (annual rates
compounded every 6 months).

• What is the current value of the forward?

45
Let d(0, k) denote the discount factor over the (0, k) semi-annual
1 1
period. We have d(0, 2) = and d(0, 1) = . Consider
(1.04)2 1.035
the future value of the cash flows associated with holding the bond
one year later and payment of F0 under the forward contract. The
current forward price of the bond
spot price c(1)d(0, 1) c(2)d(0, 2)
F0 = − −
d(0, 2) d(0, 2) d(0, 2)
2 80(1.04)2 80(1.04)2
= 920(1.04) − − 2
= 831.47.
1.035 (1.04)

The difference in the forward prices is discounted to the present

value. The current value of the forward contract = 831.47−940
(1.04)2
=
−100.34.

46
Implied forward interest rate

The forward price of a forward on a discount bond should be related

to the implied forward interest rate R(t; T1, T2). The implied forward
rate is the interest rate over the future time period [T1, T2] as implied
by time-t discount bond prices. The bond forward buyer pays F at
T1 and receives P at T2 and she is expected to earn R(t; T1, T2) over
[T1, T2], so
F [1 + R(t; T1, T2)(T2 − T1)] = P.
Together with
F = P Bt(T2)/Bt(T1),
we obtain
[ ]
1 Bt(T1)
R(t; T1, T2) = −1 .
T2 − T1 Bt(T2)

47
Forward rate agreement (FRA)

The FRA is an agreement between two counterparties to exchange

floating and fixed interest payments on the future settlement date
T2 .

• The floating rate will be the LIBOR L[T1, T2] as observed on

the future reset date T1.

Question

Should the fixed rate be set equal to the implied forward rate over
the same period (determined based on traded bond prices as ob-
served today)?

48
Determination of the forward price of LIBOR

L[T1, T2] = LIBOR to be observed at future time T1

for the accrual period [T1, T2]
K = fixed rate
N = notional of the FRA

Cash flow of the fixed rate receiver

49
Replication argument

Adding N to both parties at time T2, the cash flows of the fixed
rate payer can be replicated by

(i) long holding of the T2-maturity zero coupon bond with par N [1+
K(T2 − T1)].

(ii) short holding of the T1-maturity zero coupon bond with par N .

It is assumed that the par amount N collected at T1 will be put in

a deposit account that earns the floating LIBOR L[T1, T2].

50
Cash flow of the fixed rate receiver

51
 Value of the portfolio of bonds that replicate
the cash flow of the fixed rate receiver at the current time
= N {[1 + K(T2 − T1)]Bt(T2) − Bt(T1)}.

We find K such that the above value is zero.

[ ]
1 Bt(T1)
K= −1 .
T − T1 Bt(T2)
|2 {z }
forward rate over [T1 , T2 ]

The fair fixed rate K is visualized as the forward price of the LIBOR
rate L[T1, T2] over the time period [T1, T2].

52
1.2 Exotic swap products

Interest rate swaps

10%
Company A Company B
6-month LIBOR
Direct swap agreement

The two parties agree to exchange periodic interest payments.

• The interest payments exchanged are calculated based on some

predetermined dollar principal, called the notional amount.

• One party is the fixed-rate payer and the other party is the
floating-rate payer. The floating interest rate is based on some
reference rate (the most common index is the LONDON IN-
TERBANK OFFERED RATE, LIBOR).

53
Example

Notional amount = $50 million

fixed rate = 10%
floating rate = 6-month LIBOR

Tenor = 3 years, semi-annual payments

6-month period Cash flows
Net (float − fix) floating rate bond fixed rate bond
0 0 −50 50
1 LIBOR1 /2 × 50 − 2.5 LIBOR1 /2 × 50 −2.5
2 LIBOR2 /2 × 50 − 2.5 LIBOR2 /2 × 50 −2.5
3 LIBOR3 /2 × 50 − 2.5 LIBOR3 /2 × 50 −2.5
4 LIBOR4 /2 × 50 − 2.5 LIBOR4 /2 × 50 −2.5
5 LIBOR5 /2 × 50 − 2.5 LIBOR5 /2 × 50 −2.5
6 LIBOR6 /2 × 50 − 2.5 LIBOR6 /2 × 50 −2.5

• One interest rate swap contract can effectively establish a payoff

equivalent to a package of forward contracts.

54
A swap can be interpreted as a package of cash market instruments
– a portfolio of forward rate agreements.

• Buy $50 million par of a 3-year floating rate bond that pays
6-month LIBOR semi-annually.

• Finance the purchase by borrowing $50 million for 3 years at

10% interest rate paid semi-annually.

The fixed-rate payer has a cash market position equivalent to a

long position in a floating-rate bond and a short position in a fixed
rate bond (borrowing through issuance of a fixed rate bond). Both
bonds are assumed to be par bonds, where the present value is equal
to the par value.

55
Application to asset/liability management

• Holding a 5-year term commercial loans of $50 million with a

fixed interest rate of 10%, that is, interest of $2.5 million re-
ceived semi-annually and par received at the end of 5 years.

• To fund its loan portfolio, the bank issues 6-month certificates

of deposit with floating interest rate of LIBOR + 40 bps (100
bps = 1%). This may be a response to the market demand
from investors for floating rate certificates of deposit.

Risk 6-month LIBOR may be 9.6% or greater.

Possible strategy Swap the fixed rate income into a floating

rate cash stream to hedge against uncertainty
in floating rate.

56
Choice of swap for the bank

• Every six months, the bank will pay 8.45% (annualized rate).

• Every six month, the bank will receive LIBOR.

Outcome

To be received 10.00%+ 6-month LIBOR

To be paid 8.45% + 0.4%+ 6-month LIBOR
spread income 1.15% or 115 basis points

The bank faces potential default risk of loans.

57
Life insurance company’s position

• Has committed to pay a 9% rate for the next 5 years on a

guaranteed investment contract (GIC) of amount $50 million.

• Can invest $50 million in an attractive 5-year floating-rate in-

strument with floating interest rate of 6-month LIBOR +160
bps.

Risk 6-month LIBOR may fall to 7.4%.

Possible strategy Swap the floating rate income into a

fixed rate cash stream.

58
Choice of swap for the insurance company

• Every six months, the insurance company will pay LIBOR.

• Every six months, the insurance company will receive 8.45%.

Outcome

To be received 8.45% + 1.6%+ 6-month LIBOR

To be paid 9.00%+ 6-month LIBOR
spread income 1.05% or 105 basis points

59
Valuation of interest rate swaps

• When a swap is entered into, it typically has zero value.

• Valuation involves finding the fixed swap rate K such that the
fixed and floating legs have equal value at inception.

• Consider a swap with payment dates t1, t2, · · · , tn (tenor struc-

ture) set in the term of the swap. Li−1 is the LIBOR observed
at ti−1 but payment is made at ti. Write δi ≈ ti − ti−1 as the ac-
crual period over [ti−1, ti]. Note that δi is in general not exactly
the same as ti − ti−1 since some form of day count convention
is used to compute δi (see below).

• The fixed payment at ti is KN δi while the floating payment at

ti is Li−1N δi, i = 1, 2, · · · n. Here, N is the notional.

60
Day count convention

For the 30/360 day count convention of the time period (D1, D2]
with D1 excluded but D2 included, the year fraction is
max(30 − d1, 0) + min(d2, 30) + 360 × (y2 − y1) + 30 × (m2 − m1 − 1)
360
where di, mi and yi represent the day, month and year of date Di, i =
1, 2.

For example, the year fraction between Feb 27, 2006 and July 31,
2008
30 − 27 + 30 + 360 × (2008 − 2006) + 30 × (7 − 2 − 1)
=
360
33 4
= +2+ .
360 12

61
Replication of cash flows

The fixed payment at Ti is KN δi. The fixed payments are packages

of discount bonds with par KN δi at maturity date Ti, i = 1, 2, · · · , n.

To replicate the floating leg payments at current time t, t < T0, we

long T0-maturity discount bond with par N and short Tn-maturity
discount bond with par N . The N dollars collected at T0 can gen-
erate the floating leg payments Li−1N δi at all Ti by rolling over N
dollars in a deposit bank account earning interest rate Li−1 over
[Ti−1, Ti], i = 1, 2, · · · , n. Note that we always have N dollars at the
beginning of each accrual period since the interests earned are used
to honor the floating leg payments. The remaining N dollars at Tn
is used to pay the par of the Tn-maturity bond.

By paying N B(t, T0) at time t to acquire the T0-maturity discount

bond, we can generate these floating leg payments at all time points
and the par N at tn.

62
 • Let B(t, T ) be the time-t value of the discount bond with ma-
turity t.

Present value of floating leg payments = N [B(t, T0) − B(t, Tn)]

n
∑
Sum of present value of fixed leg payments = N K δi B(t, Ti).
i=1

The swap rate K is given by equating the present values of the two
sets of payments:
B(t, T0) − B(t, Tn)
K= ∑n .
i=1 δiB(t, Ti )

The interest rate swap reduces to a FRA when n = 1. As a check,

we obtain
B(t, T0) − B(t, T1)
K= ,
(T1 − T0)B(t, T1)
just the same formula as shown on p.52.

63
Asset swap

• Combination of a defaultable bond with an interest rate swap.

B pays the notional amount upfront to acquire the asset swap
package.

1. A fixed defaultable coupon bond issued by C with coupon c

payable on coupon dates.

2. A fixed-for-floating swap.

64
Remarks

1. Asset swap transactions are driven by the desire to replace the

fixed coupons by floating coupons. Asset swaps are more liquid
than the underlying defaultable bond.

2. By market convention, the whole package is sold at par. The

asset swap spread sA is adjusted to ensure that the asset swap
package has an initial value equal to the par value of the de-
faultable bond.

For convenience of our discussion, we set the par to be unity.

65
Default free bond price

C(t) = time-t price of default-free bond with fixed-coupon c

B(t, T ) = time-t price of default-free zero-coupon bond with unit par

Defaultable bond price

C(t) = time-t price of defaultable bond with fixed-coupon c

The difference C(t) − C(t) represents the credit risk premium of the
defaultable bond. Investors pay a lesser amount of C(t) − C(t) due
to potential losses arising from bond default.

66
Forward swap rate

Recall that the swap rate is the fixed rate in an interest rate swap
where the floating rate payer pays LIBOR and receives the swap
rate (fixed). There is always a tenor structure that underlies an
interest rate swap.

s(t) = forward swap rate at time t of a standard fixed-for-floating

over the tenor [tn, tn+1, . . . , tN ], t ≤ tn .
The first swap payment starts on tn+1 (based on the accrual period
[tn, tn+1]) and the last payment date is tN .

The theoretical forward rate can be determined in terms of discount

bond prices based on replication (see p.63). However, interest rate
swaps are highly liquid instruments and the forward swap rates are
market observable. It may occur that the market swap rate may not
agree exactly with the above theoretical formula based on traded
bond prices.

67
 Asset swap packages

An asset swap package consists of a defaultable coupon bond C

with coupon c and an interest rate swap. The bond’s coupon is
swapped into LIBOR plus the asset swap rate sA.

Payoff streams to the buyer of the asset swap package

time defaultable bond interest rate swap net

t = 0† −C(0) −1 + C(0) −1
t = ti c∗ −c + Li−1 + sA Li−1 + sA + (c∗ − c)
t = tN (1 + c)∗ −c + LN −1 + sA 1∗ + LN −1 + sA + (c∗ − c)
default recovery unaffected∗∗ recovery

⋆ denotes payment contingent on survival.

† The value of the interest rate swap at t = 0 is not zero. The sum of the
values of the interest rate swap and defaultable bond is set to be equal to
par at t = 0 based on market convention.

∗∗ The interest rate swap continues even after the bond defaults.

68
Remarks

• We do not need to model uncertainties in the coupons paid by

the defaultable bond since the market has evaluated the expec-
tation of the potential losses to the holder through the market
price C(t) of the defaultable bond. Luckily, we are able to put
all dusts (uncertainties of the time of arrival of default and re-
covery value) under the carpet (market price of the defaultable
bond).

• However, if the interest rate swap discontinues after default of

the underlying bond, we need to model the random time of
arrival of default in the pricing procedure since there will be no
exchange of interest payments after bond default.

69
The additional asset spread sA above LIBOR serves as the compen-
sation for bearing the potential loss upon default.

s(0) = fixed-for-floating swap rate (market quote)

A(0) = value of an annuity paying at $1 per annum over the same

tenor as the interest rate swap (calculated based on observable
default free bond prices)

Suppose the annuity stream of $1 per annum makes payments quar-

terly, this would mean $0.25 to be paid in each quarterly payment,
making a total of $0.25 × 4 = $1 per annum. Once the tenor is
known, we can find the value of the annuity for a given tenor as
sum of present values of the cash flows generated from the annuity.

Now, suppose the fixed coupon per annum is 3%, paid quarterly.
Since the par is assumed to be unity, then each coupon payment is
$0.03/4 = $0.0075. The value of the stream of the fixed coupons
for the given tenor is 0.03A(0).

70
The value of asset swap package is set at par at t = 0, so that

C(0) + A(0)s(0) + A(0)sA (0) − A(0)c = 1. (A)

| {z }
swap arrangement
The present value of the floating coupons is given by A(0)s(0). This
is because the value of the floating leg payments at LIBOR is the
same as the value of the fixed leg payments at the market swap
rate s(0) under the market interest rate swap. It has been assumed
that the interest rate swap continues even after default so that A(0)
appears in all terms associated with the swap arrangement.

Solving for sA(0), we obtain the asset spread

1
sA(0) = [1 − C(0)] + c − s(0).
A(0)

71
Rearranging the terms in eq.(A), we obtain

C(0) + A(0)sA(0) = [1 |
− A(0)s(0)]
{z
+ A(0)c} ≡ C(0)
default-free bond
where the right-hand side gives the value of a default-free bond
with coupon c. A default free bond that pays LIBOR is called a
par floater since its fair value at initiation equals par. Stripping all
floating leg payments that pay LIBOR reduces the par floater to
single payment of $1 at maturity tN . Therefore, 1 − A(0)s(0) is the
present value of receiving $1 at maturity tN . We obtain
1
sA(0) = [C(0) − C(0)].
A(0)
The extra asset swap sA accounts for the potential loss upon default.
This is quantified as C(0) − C(0) based on market bond prices at
t = 0. For a fair deal between the two parties, the difference in
the bond prices should be set equal to the present value of annuity
stream at the rate sA(0).

72
Alternative proof

A combination of the non-defaultable counterpart (bond with coupon

rate c) plus an interest rate swap (whose floating leg is LIBOR while
the fixed leg is c) becomes a par floater. Hence, the new asset pack-
age should also be sold at par.

The buyer is guaranteed to receive LIBOR floating rate interests

plus par. We have
PV $1 at tN
z }| {
C(0) = 1 − A(0)s(0) +A(0)c
while A(0)s(0) − A(0)c gives the value of the interest rate swap.
73
In other words, we obtain

value of interest rate swap + price of nondefaultable bond

= [A(0)s(0) − A(0)c] + [1 − A(0)s(0) + A(0)c] = 1.

The two interest swaps with floating leg at LIBOR + sA(0) and
LIBOR, respectively, differ in values by sA(0)A(0).

Comparing the two asset swaps on P.64 and P.73, the earlier one
has defaultable bond while the later one has default free bond. The
price difference is C(0) − C(0). This is compensated by the present
value of the annuity at the rate sA(0) per annum. Therefore,

C(0) − C(0) = sA(0)A(0).

74
Total return swap

• Exchange the total economic performance of a specific asset for

another cash flow.
total return of asset
Total return Total return
payer receiver
LIBOR + Y bp

Total return comprises the sum of interests, fees and any

change-in-value payments with respect to the reference asset.

A commercial bank can hedge all credit risk on a bond/loan it has

originated. The counterparty can gain access to the bond/loan on
an off-balance sheet basis, without bearing the cost of originating,
buying and administering the loan. The TRS terminates upon the
default of the underlying asset.

75
Used as a financing tool

• The receiver wants financing to invest $100 million in the refer-

ence bond. It approaches the payer (a financial institution) and
agrees to the swap.
• The payer invests $100 million in the bond. The payer retains
ownership of the bond for the life of the swap and has much
less exposure to the risk of the receiver defaulting (as compared
to the actual loan of $100 million).
• The receiver is in the same position as it would have been if it
had borrowed money at LIBOR + sT RS to buy the bond. He
bears the market risk and default risk of the underlying bond.

76
Some essential features

1. The receiver is synthetically long the reference asset without

having to fund the investment up front. He has almost the
same payoff stream as if he had invested in risky bond directly
and funded this investment at LIBOR + sT RS .
2. The TRS is marked to market at regular intervals, similar to a
futures contract on the risky bond. The reference asset should
be liquidly traded to ensure objective market prices for marking
to market (determined using a dealer poll mechanism).
3. The TRS allows the receiver to leverage his position much higher
than he would otherwise be able to (may require collateral). The
TRS spread should not only be driven by the default risk of the
underlying asset but also by the credit quality of the receiver.

77
The payments received by the total return receiver are:

1. The coupon c of the bond (if there were one since the last
payment date Ti−1).
2. The price appreciation (C(Ti)−C(Ti−1))+ of the underlying bond
C since the last payment (if there were any).
3. The recovery value of the bond (if there were default).

78
The payments made by the total return receiver are:

1. A regular fee of LIBOR +sT RS .

2. The price depreciation (C(Ti−1) − C(Ti))+ of bond C since the
last payment (if there were any).
3. The par value of the bond C (if there were a default in the
meantime).

The coupon payments are netted and swap’s termination date is

earlier than bond’s maturity.

79
Motivation of the receiver

1. Investors can create new assets with a specific maturity not

currently available in the market.
2. Investors gain efficient off-balance sheet exposure to a desired
asset class to which they otherwise would not have access.
3. Investors may achieve a higher leverage on capital – ideal for
hedge funds. Otherwise, direct asset ownership is on on-balance
sheet funded investment.
4. Investors can reduce administrative costs via an off-balance sheet
purchase.
5. Investors can access entire asset classes by receiving the total
return on an index.

80
Motivation of the payer

• A long-term investor, who feels that a reference asset in the

portfolio may widen in spread in the short term but will recover
later, may enter into a total return swap that is shorter than the
maturity of the asset. She can gain from the price depreciation.
This structure is flexible and does not require a sale of the asset
(thus accommodates a temporary short-term negative view on
an asset).

81
Credit default swaps

The protection seller receives fixed periodic payments from the pro-
tection buyer in return for making a single contingent payment cov-
ering losses on a reference asset following a default.
140 bp per annum
protection protection
seller buyer
Credit event payment
(100%  recovery rate)
only if credit event occurs

holding a
risky bond

82
Protection seller

• earns premium income with no funding cost

• gains customized (tailored made expiration date) and, synthetic

(without physical holding) access to the risky bond

Protection buyer

• hedges the default risk on the reference asset

1. Very often, the bond tenor is longer than the swap tenor. In
this way, the protection seller does not have exposure to the full
period of the bond.

2. Basket default swap – gain additional yield by selling default

protection on several assets.

83
A bank lends 10mm to a corporate client at L + 65bps. The bank
also buys 10mm default protection on the corporate loan for 50bps.

Objective achieved for the bank

• maintain relationship with the corporate borrower

••transfer the credit risk on a new loan

Risk Transfer

Default Swap
Premium
Corporate Interest and Financial
Bank If Credit Event:
Borrower Principal House
par amount
If Credit Event:
obligation (loan)

84
Settlement of compensation payment

1. Physical settlement:
The defaultable bond is put to the Protection Seller in return
for the par value of the bond.

2. Cash compensation:
An independent third party determines the loss upon default
at the end of the settlement period (say, 3 months after the
occurrence of the credit event).

Compensation amount = (1 − recovery rate) × bond par.

85
Selling protection

To receive credit exposure for a fee or in exchange for credit expo-

sure to better diversify the credit portfolio.

Buying protection

To reduce either individual credit exposures or credit concentrations

in portfolios. Synthetically to take a short position in an asset
which are not desired to sell outright, perhaps for relationship or
tax reasons.

86
The price of a corporate bond must reflect not only the spot rates
for default-free bonds but also a risk premium to reflect default risk
and any options embedded in the issue.

Credit spreads: compensate investor for the risk of default on the

underlying securities

Construction of a credit risk adjusted yield curve is hindered by

1. The general absence in money markets of liquid traded instru-

ments on credit spread. For liquidly traded corporate bonds,
we may have good liquidity on trading of credit default swaps
whose underlying is the credit spread.

2. The absence of a complete term structure of credit spreads as

implied from traded corporate bonds. At best we only have
infrequent data points.

87
• The spread increases as the rating declines. It also increases
with maturity.

• The spread tends to increase faster with maturity for low credit
ratings than for high credit ratings.

88
Funding cost arbitrage

Should the Protection Buyer look for a Protection Seller who has a
higher/lower credit rating than himself?

A-rated institution 50bps AAA-rated institution LIBOR-15bps Lender to the

as Protection Seller annual as funding AAA-rated
as Protection Buyer
premium cost Institution
funding cost of coupon
LIBOR + 50bps = LIBOR + 90bps

Lender to the BBB risky

A-rated Institution reference asset

89
The combined risk faced by the Protection Buyer:

• default of the BBB-rated bond

• default of the Protection Seller on the contingent payment

Consider the S&P’s Ratings for jointly supported obligations (the

two credit assets are uncorrelated)

A+ A A− BBB+ BBB
A+ AA+ AA+ AA+ AA AA
A AA+ AA AA AA− AA−

The AAA-rated Protection Buyer creates a synthetic AA−asset with

a coupon rate of LIBOR + 90bps − 50bps = LIBOR + 40bps.
This is better than LIBOR + 30bps, which is the coupon rate of a
AA−asset (net gains of 10bps).

90
For the A-rated Protection Seller, it gains synthetic access to a
BBB-rated asset with earning of net spread of

• Funding cost of the A-rated Protection Seller = LIBOR +50bps

• Coupon from the underlying BBB bond = LIBOR +90bps

• Credit swap premium earned = 50bps

91
In order that the credit arbitrage works, the funding cost of the
default protection seller must be higher than that of the default
protection buyer.

Example

Suppose the A-rated institution is the Protection Buyer, and assume

that it has to pay 60bps for the credit default swap premium (higher
premium since the AAA-rated institution has lower counterparty
risk).

spread earned from holding the risky bond

= coupon from bond − funding cost
= (LIBOR + 90bps) − (LIBOR + 50bps) = 40bps
which is lower than the credit swap premium of 60bps paid for
hedging the credit exposure. No deal is done!

92
Counterparty risk in CDS

Before the Fall 1997 crisis, several Korean banks were willing to
offer credit default protection on other Korean firms.

40 bp
US commercial Korea exchange
bank bank

LIBOR + 70bp

Hyundai
(not rated)

⋆ Higher geographic risks lead to higher default correlations.

Advice: Go for a European bank to buy the protection.

93
How does the inter-dependent default risk structure between the
Protection Seller and the Reference Obligor affect the swap rate?

1. Replacement cost (Seller defaults earlier)

• If the Protection Seller defaults prior to the Reference En-

tity, then the Protection Buyer renews the CDS with a new
counterparty.

• Supposing that the default risks of the Protection Seller and

Reference Entity are positively correlated, then there will be
an increase in the swap rate of the new CDS.

2. Settlement risk (Reference Entity defaults earlier)

• The Protection Seller defaults during the settlement period

after the default of the Reference Entity.

94
Fixed-coupon bonds

Portfolio 1

• One defaultable coupon bond C; coupon c, maturity tN .

• One CDS on this bond, with CDS spread s

The portfolio is unwound after a default.

Portfolio 2

• One default-free coupon bond C: with the same payment dates

as the defaultable coupon bond and coupon size c − s.

The bond is sold after default.

95
Comparison of cash flows of the two portfolios

1. In survival, the cash flows of both portfolio are identical.

Portfolio 1 Portfolio 2
t=0 −C(0) −C(0)
t = ti c−s c−s
t = tN 1+c−s 1+c−s

2. At default, portfolio 1’s value = par = 1 (full compensation by

the CDS); that of portfolio 2 is C(τ ), τ is the time of default.
The price difference at default = 1 − C(τ ). This difference is
very small when the default-free bond is a par bond.

Remark

The issuer can choose c to make the bond be a par bond such that
the initial value of the bond is at par.
96
This is an approximate replication.

Recall that the value of the CDS at time 0 is zero since it costs
nothing on both sides to enter into a swap. Neglecting the differ-
ence in the values of the two portfolios at default, the no-arbitrage
principle dictates

C(0) = C(0) = B(0, tN ) + cA(0) − sA(0).

Here, B(0, tN ) is the time-0 price of a unit par default free discount
bond maturing at tN and (c − s)A(0) is the sum of present value of
the coupon payments at the fixed coupon rate c−s. The equilibrium
CDS rate s can be solved:
B(0, tN ) + cA(0) − C(0) C(0) − C(0)
s= = ,
A(0) A(0)
where B(0, tN ) + cA(0) = C(0) is the time-0 price of a default free
coupon bond paying coupon at the rate of c.

It is not surprising to observe the analogy between the asset swap

spread and CDS premium.
97
Floating rate bonds
′
A par floater C is a defaultable bond with a floating-rate coupon
of ci = Li−1 + spar , where the par spread spar is adjusted such that
at issuance the par floater is valued at par.

Portfolio 1

′
• One defaultable par floater C with spread spar over LIBOR.
• One CDS on this bond: CDS spread is s.

The portfolio is unwound after default.

98
Portfolio 2

• One default-free floating-coupon bond C ′: with the same pay-

ment dates as the defaultable par floater and coupon at LIBOR,
ci = Li−1.
The bond is sold after default.
Time Portfolio 1 Portfolio 2
t=0 −1 −1
t = ti Li−1 + spar − s Li−1
t = tN 1 + LN −1 + spar − s 1 + LN −1
τ (default) 1 C ′(τ ) = 1 + Li (τ − ti )

The hedge error in the payoff at default is caused by accrued interest

Li(τ − ti), accumulated from the last coupon payment date ti to the
default time τ . If we neglect the small hedge error at default, if we
set
spar = s,
then the portfolios are seen to be essentially equivalent.
99
Remarks

• The non-defaultable bond becomes a par bond (with initial value

equals the par value) when it pays the floating rate equals LI-
BOR. The extra coupon spar paid by the defaultable par floater
represents the credit spread demanded by the investor due to
the potential credit risk. The above result shows that the credit
spread spar is just equal to the CDS spread s.

• The above analysis neglects the counterparty risk of the Pro-

tection Seller of the CDS. Due to potential counterparty risk,
the actual CDS spread will be lower.

100
IBM/World Bank – first currency swap structured in 1981

• IBM had existing debts in DM and Swiss francs. This had creat-
ed a foreign exchange exposure since IBM had to convert USD
into DM and Swiss Francs regularly to make the coupon pay-
ments. Due to a depreciation of the DM and Swiss franc against
the dollar, IBM could realize a large foreign exchange gain, but
only if it could eliminate its DM and Swiss franc liabilities and
“lock in” the gain and remove any future exposure.

• The World Bank was raising most of its funds in DM (interest

rate = 12%) and Swiss francs (interest rate = 8%). It did not
borrow in dollars, for which the interest rate cost was about
17%. Though it wanted to lend out in DM and Swiss francs,
the bank was concerned that saturation in the bond markets
could make it difficult to borrow more in these two currencies
at a favorable rate. Its objective, as always, was to raise cheap
funds.

101
102
• IBM was willing to take on dollar liabilities and made dollar
payments (periodic coupons and principal at maturity) to the
World Bank since it could generate dollar income from its normal
trading activities.

• The World Bank could borrow dollars, convert them into DM

and SFr in FX market, and through the swap take on payment
obligations in DM and SFr.

1. The foreign exchange gain on dollar appreciation is realized by

IBM through the negotiation of a favorable swap rate in the
currency swap contract.

2. The swap payments by the World Bank to IBM were scheduled

so as to allow IBM to meet its debt obligations in DM and SFr.

103
Under the currency swap

• IBM pays regular US coupons and US principal at maturity.

• World Bank pays regular DM and SFr coupons together with

DM and SFr principal at maturity.

Now IBM converted its DM and SFr liabilities into USD, and the
World Bank effectively raised hard currencies at a cheap rate. Both
parties achieved their objectives!

104
Differential Swap (Quanto Swap)

A special type of floating-against-floating currency swap that does

not involve any exchange of principal, not even at maturity.

• Interest payments are exchanged by reference to a floating rate

index in one currency and another floating rate index in a second
currency. Both interest rates are applied to the same notional
principal in one currency.
• Interest payments are made in the same currency.

Apparently, the risk factors are a floating domestic interest rate

and a floating foreign interest rate. However, since foreign floating
rates are applied on domestic payments, the correlation between
exchange rate and foreign floating rate poses correlation risk.

105
All cashflows are denominated in the same currency.
106
Rationale

To exploit large differential in short-term interest rates across major

currencies without directly incurring exchange rate risk.

Applications

• Money market investors use diff swaps to take advantage of a

high yield currency if they expect yields to persist in a discount
currency.
• Corporate borrowers with debt in a discount currency can use
diff swaps to lower their effective borrowing costs from the ex-
pected persistence of a low nominal interest rate in the premium
currency. Pay out the lower floating rate in the premium curren-
cy in exchange to receive the high floating rate in the discount
currency.

107
 • The value of a diff swap in general would not be zero at initia-
tion. The value is settled either as an upfront premium payment
or amortized over the whole life as a margin over the floating
rate index.

Uses of a differential swap

Suppose a company has hedged its liabilities with a dollar interest

rate swap serving as the fixed rate payer, the shape of the yield
curve in that currency will result in substantial extra costs. The cost
is represented by the differential between the short-term 6-month
dollar LIBOR and medium to long-term implied LIBORs payable in
dollars – upward sloping yield curve.

108
 • The borrower enters into a dollar interest rate swap whereby it
pays a fixed rate and receives a floating rate (6-month dollar
LIBOR).

• Simultaneously, it enters into a diff swap for the same dollar

notional principal amount whereby the borrower agrees to pay
6-month dollar LIBOR and receive 6-month Euro LIBOR less a
margin.

The result is to increase the floating rate receipts under the dollar
interest rate swap so long as 6-monthly Euro LIBOR, adjusted for
the diff swap margin, exceeds 6-month LIBOR. This has the impact
of lowering the effective fixed rate cost to the borrower.

109
The borrower has been forced to pay a high fixed rate of 7.25% due
to the upward sloping yield curve of LIBOR. On the other hand,
this may help the borrower to obtain a lower margin. The borrower
gains if the upward trend of LIBOR is not realized.

110
1.3 Options: Rational boundaries of option values

Financial options

• A call (or put) option is a contract which gives the holder the
right to buy (or sell) a prescribed asset, known as the underlying
asset, by a certain date (expiration date) for a predetermined
price (commonly called the strike price or exercise price).

• The option is said to be exercised when the holder chooses to

buy or sell the asset.

• If the option can only be exercised on the expiration date, then

the option is called a European option.

• If the exercise is allowed at any time prior to the expiration date,

then it is called an American option

111
Terminal payoff

• Let ST denote the asset price at maturity date T and X be the

strike price.
• The terminal payoff from the long position (holder’s position)
of a European call is then

max(ST − X, 0).

• The terminal payoff from the long position in a European put

can be shown to be

max(X − ST , 0),
since the put will be exercised at expiry only if ST < X, whereby
the asset worth ST is sold at a higher price of X.

112
Note that call’s terminal payoff minus put’s terminal payoff equals
the terminal payoff of a forward contract: ST − X.

113
Questions and observations

What should be the fair option premium (usually called option price
or option value) so that the deal is fair to both writer and holder?

What should be the optimal strategy to exercise prior to the expi-

ration date for an American option?

At least, the option price is easily seen to depend on the strike price,
time to expiry and current asset price. The less obvious factors for
the pricing models are the prevailing interest rate and the degree of
randomness of the asset price, commonly called the volatility .

114
Hedging a short position in a call

• If the writer of a call does not simultaneously own any amount

of the underlying asset, then he is said to be in a naked position.

• Suppose the call writer owns some amount of the underlying

asset, the loss in the short position of the call when asset price
rises can be compensated by the gain in the long position of the
underlying asset.

• This strategy is called hedging, where the risk in a portfolio is

monitored by taking opposite directions in two securities which
are highly negatively correlated.

• In a perfect hedge situation, the hedger combines a risky op-

tion and the corresponding underlying asset in an appropriate
proportion to form a riskless portfolio.

115
Swaptions – Product nature

• The buyer of a swaption has the right to enter into an interest

rate swap by some specified date. The swaption also specifies
the maturity date of the swap. The buyer of the swaption pays
the premium upfront.

• The buyer can be the fixed-rate receiver or the fixed-rate payer.

The writer becomes the counterparty to the swap if the buyer
exercises. The strike rate indicates the fixed rate that will be
swapped versus the floating rate.

• Suppose the buyer of the swaption is the fixed rate payer in the
underlying swap, she chooses to exercise the swaption when the
prevailing swap rate on swaption’s maturity date is higher than
the strike rate. This is because the swaption buyer would pay
a lower fixed rate in the interest rate swap under the swaption
contract when compared with the higher prevailing fixed rate in
a newly negotiated interest rate swap.

116
• To the buyer (fixed rate payer), the exercise of the swaption is
the sale of the fixed rate bond at the price of the floating rate
bond.
• The value of the floating rate bond equals the par at initiation of
the swap, so it may be viewed as having fixed value (strike price
of the swaption). This swaption is thus seen as a put swaption
since the holder sells a fixed rate bond with floating value to re-
ceive a floating rate bond with fixed value upon exercise. Recall
that the holder of a put option has the right to forfeit an asset
with floating value to receive fixed amount of dollars.
• Note that the value of the fixed rate bond with coupon rate same
as the prevailing swap rate would be equal to the par value.
Obviously, the value of the fixed rate bond with coupon rate
same as the preset strike rate in the swaption (lower than the
prevailing swap rate) would be below par. When the swaption
(fixed rate payer) holder exercises the swaption, the floating
value asset (fixed rate bond) is below the strike price (par value
of the floating rate bond) of the put swaption.

117
Uses of swaptions

Used to hedge a portfolio strategy that involves the use of an interest

rate swap while the cash flow of the underlying asset or liability is
uncertain. Uncertainties come from (i) early termination of cash
flows due to callability in a callable bond or prepayment of mortgage
loans, (ii) exposure to default risk.

Example

Consider a Savings & Loans Association entering into a 4-year swap

in which it agrees to pay 9% fixed and receive LIBOR.

• The fixed rate payments come from a portfolio of mortgage

pass-through securities with a coupon rate of 9%. One year
later, mortgage rates decline, resulting in large prepayments.

• The purchase of a call swaption with a strike rate of 9% would

be useful to offset the original swap position.

118
Due to decline in the interest rate, large prepayments are resulted
in the mortgage pass-through securities. The source of 9% fixed
payment dissipates. The swaption is in-the-money since the interest
rate declines, so does the swap rate.

119
By exercising the call swaption, the Savings & Loans Association
receives a fixed rate of 9%. The risk of potential decline in interest
rates is hedged via the purchase of a swaption.

• Treating the fixed rate bond as the underlying asset in the swap-
tion and the floating rate bond as the fixed par value, the “pay-
float” swaption is visualized as a call swaption since the holder
pays the fixed strike to receive the underlying asset.

120
Management of callable debts

Three years ago, XYZ issued 15-year fixed rate callable debt with a
coupon rate of 12%.

Strategy

The bond issuer XYZ sells a two-year fixed rate receiver option on a
10-year swap that gives the holder the right but not the obligation
to receive the fixed rate of 12%.

121
Call monetization

By selling the European call swaption with two-year maturity today,

the company has committed itself to paying fixed 12% coupon for
the remaining life of the original bond.

• The call swaption was sold in exchange for an upfront swaption

premium received at date 0. The monetization of the callable
right is realized via the swaption premium received.

122
Call-Monetization cash flow: Swaption expiration date

Interest rate (swap rate) ≥ 12%

Continue to pay
12% coupon

• The call swaption holder (fixed rate receiver) does not exercise
the European swaption on expiration date since he would prefer
to receive the prevailing swap rate instead of strike rate of 12%.

• XYZ earns the full proceed of the swaption premium

123
 Interest rate (swap rate) < 12%

• Counterparty of the European swaption exercises the swaption

• XYZ calls the bond since the prevailing swap rate is below the
promised 12% coupon rate in the bond. Once the old bond is
retired, XYZ issues a new floating rate bond that pays floating
rate LIBOR (funding rate would also depend on the creditwor-
thiness of XYZ at that time). The cash required to retire the old
fixed rate bond is financed by the issuance of the new floating
rate bond.
124
Rational boundaries for option values

• Mathematical properties of the option values as functions of the

strike price X, asset price S and time to expiry τ are derived. We
do not specify the probability distribution of the movement of
the asset price so that the fair option value cannot be derived.
We study the impact of dividends on these rational boundaries.

• The optimal early exercise policy of American options on a non-

dividend paying asset can be inferred from the analysis of these
bounds on option values.

• The relations between put and call prices (called the put-call
parity relations) are also deduced. These relations are distribu-
tion free.

125
Non-negativity of option prices

All option prices are non-negative, that is,

C ≥ 0, P ≥ 0, c ≥ 0, p ≥ 0,
as derived from the non-negativity of the payoff structure of option
contracts.

If the price of an option were negative, this would mean an option

buyer receives cash up-front. He is guaranteed to have a non-
negative terminal payoff. In this way, he can always lock in a riskless
profit.

126
Intrinsic values of American options

• max(S−X, 0) and max(X−S, 0) are commonly called the intrinsic

value of a call and a put, respectively.

• Since American options can be exercised at any time before

expiration, their values must be worth at least their intrinsic
values, that is,

C(S, τ ; X) ≥ max(S − X, 0)
P (S, τ ; X) ≥ max(X − S, 0).

• Suppose C is less than S−X when S ≥ X, then an arbitrageur can

lock in a riskless profit by borrowing C + X dollars to purchase
the call and exercise it immediately to receive the asset worth
S. The riskless profit would be S − X − C > 0. Market frictions
may lead to violation of the lower bound on call value (like near-
maturity warrants).

127
American options are worth at least their European counterparts

An American option confers all the rights of its European coun-

terpart plus the privilege of early exercise. The additional privilege
cannot have negative value.

C(S, τ ; X) ≥ c(S, τ ; X)
P (S, τ ; X) ≥ p(S, τ ; X).

• The European put value can be below the intrinsic value X − S

at sufficiently low asset value. In this case, the European put is
almost sure to expire in-the-money. However, the par is received
only at maturity for the European put.

• The value of a European call on a dividend paying asset can be

below the intrinsic value S − X at sufficiently high asset value.
This is because holding of the European call does not entitle
the holder to receive the dividends while the asset price drops in
value when sizable dividends are paid during the life of the call.

128
Values of options with different dates of expiration

Consider two American options with different times to expiration τ2

and τ1 (τ2 > τ1), the one with the longer time to expiration must
be worth at least that of the shorter-lived counterpart since the
longer-lived option has the additional right to exercise between the
two expiration dates.

C(S, τ2; X) > C(S, τ1; X), τ2 > τ1,

P (S, τ2; X) > P (S, τ1; X), τ2 > τ1 .
The above argument cannot be applied to European options due to
lack of the early exercise privilege.

129
Values of options with different strike prices

c(S, τ ; X2) < c(S, τ ; X1), X1 < X2,

C(S, τ ; X2) < C(S, τ ; X1), X1 < X 2 .
and

p(S, τ ; X2) > p(S, τ ; X1), X1 < X2,

P (S, τ ; X2) > P (S, τ ; X1), X 1 < X2 .

Values of options at varying asset value levels

c(S2, τ ; X) > c(S1, τ ; X), S 2 > S1 ,

C(S2, τ ; X) > C(S1, τ ; X), S2 > S1;
and

p(S2, τ ; X) < p(S1, τ ; X), S2 > S1,

P (S2, τ ; X) < P (S1, τ ; X), S2 > S1.

130
Upper bounds on call and put values

• A call option is said to be a perpetual call if its date of expiration

is infinitely far away. The asset itself can be considered as an
American perpetual call with zero strike price plus additional
privileges such as voting rights and receipt of dividends, so we
deduce that S ≥ C(S, ∞; 0).

S ≥ C(S, ∞; 0) ≥ C(S, τ ; X) ≥ c(S, τ ; X).

• The price of an American put equals the strike value when the
asset value is zero; otherwise, it is bounded above by the strike
price.
X ≥ P (S, τ ; X) ≥ p(S, τ ; X).

131
Lower bounds on the values of call options on a non-dividend paying
asset

Portfolio A consists of a European call on a non-dividend paying

asset plus a discount bond with a par value of X whose date of
maturity coincides with the expiration date of the call. Portfolio B
contains one unit of the underlying asset.

Asset value at expiry ST < X ST ≥ X

Portfolio A X (ST − X) + X = ST
Portfolio B ST ST
Result of comparison VA > VB VA = VB

The present value of Portfolio A (dominant portfolio) must be equal

to or greater than that of Portfolio B (dominated portfolio). If oth-
erwise, an arbitrage opportunity can be secured by buying Portfolio
A and selling Portfolio B. This is called the principle of dominance
(generalization of the law of one price).

132
Write B(τ ) as the price of the unit-par discount bond with time to
expiry τ . Then
c(S, τ ; X) + XB(τ ) ≥ S.
Together with the non-negativity property of option value.

c(S, τ ; X) ≥ max(S − XB(τ ), 0).

The upper and lower bounds of the value of a European call on a
non-dividend paying asset are given by (see Figure)

S ≥ c(S, τ ; X) ≥ max(S − XB(τ ), 0).

Remark
Earlier, we have established C(S, τ ; X) ≥ max(S − X, 0) for an Amer-
ican call option. Note that C(S, τ ; X) ≥ c(S, τ ; X) and max(S −
XB(τ ), 0) ≥ max(S − X, 0), so the lower bound for a European call
max(S − XB(τ ), 0) is even sharper.

133
134
Convexity properties of the option price functions

The call prices are convex functions of the strike price. Write X2 =
λX3 + (1 − λ)X1, where 0 ≤ λ ≤ 1, X1 ≤ X2 ≤ X3, the convexity
property gives

c(S, τ ; X2) ≤ λc(S, τ ; X3) + (1 − λ)c(S, τ ; X1)

135
 • The drop in the European call value for one dollar increase in
the strike price should be less than one dollar. The loss in
the terminal payoff of the call due to the increase in the strike
price is realized only
 when
 the call expires in-the-money. More
 ∂c 
 
precisely, we have   < B(τ ). The factor B(τ ) appears since
 ∂X 
the potential loss of paying extra one dollar in the strike price
occurs at maturity so its present value is B(τ ).

Consider the payoffs of the following two portfolios at expiry. Port-

folio C contains λ units of call with strike price X3 and (1 − λ) units
of call with strike price X1, and Portfolio D contains one call with
strike price X2.

Since VC ≥ VD for all possible values of ST , Portfolio C is dominant

over Portfolio D. Therefore, the present value of Portfolio C must
be equal to or greater than that of Portfolio D.

136
Payoff at expiry of Portfolios C and D.

Asset value ST ≤ X1 X1 ≤ ST ≤ X2 X2 ≤ ST ≤ X3 X3 ≤ ST
at expiry
Portfolio C 0 (1 − λ)(ST − X1 ) (1 − λ)(ST − X1 ) λ(ST − X3 ) +
(1 − λ)(ST − X1 )
Portfolio D 0 0 ST − X2 ST − X2
Result of VC = VD VC ≥ VD VC ≥ VD∗ VC = VD
comparison

* Recall X2 = λX3 + (1 − λ)X1, and observe

(1 − λ)(ST − X1) ≥ ST − X2
⇔ X2 − (1 − λ)X1 ≥ λST
⇔ X3 ≥ ST .

137
 • By changing the call options in the above two portfolios to the
corresponding put options, it can be shown that European put
prices are also convex functions of the strike price.

• By using the linear homogeneity property of the call and put

option functions with respect to the asset price and strike price,
where for the case of an European call option:

c(λS, τ, λX) = λc(S, τ, X), λ > 0,

one can show that the European call prices are convex functions
of the asset price. That is,

c(λS1 + (1 − λ)S2, τ ; X) ≤ λc(S1, τ, X) + (1 − λ)c(S2, τ ; X), λ > 0.

Note that we first establish convexity property of option price func-

tions on strike via the dominance principle since we can hold port-
folios of call options with different strikes and same expiration date.
We then use the linear homogeneity property to establish convexity
property on asset price.
138
Impact of dividends on the asset price

• When an asset pays a certain amount of dividend, no arbitrage

argument dictates that the asset price falls by the same amoun-
t (assuming there exist no other factors affecting the income
proceeds, like taxation and transaction costs).

• Suppose the asset price falls by an amount less than the dividend,
an arbitrageur can lock in a riskless profit by borrowing money
to buy the asset right before the dividend date, selling the asset
right after the dividend payment and returning the loan.

139
It is assumed that the deterministic dividend amount Di is paid at
time ti, i = 1, 2, · · · , n. The current time is t and write τi = ti − t, i =
1, 2, · · · , n. The sum of the present value of the dividends within the
life of the option is

D = D1e−rτ1 + · · · + Dne−rτn .
The dividends stream may be visualized as a portfolio of bonds with
par value Di maturing at ti, i = 1, 2, · · · , n.

Weakness in the assumption

One may query whether the asset can honor the deterministic divi-
dend payments when the asset value becomes very low.

140
Put-call parity relations

For a pair of European put and call on the same underlying asset
and with the same expiration date and strike price, we have

p = c − S + D + XB(τ ).
When the underlying asset is non-dividend paying, we set D = 0.
Recall that the value of the forward f = S − D − XB(τ ), so c − p = f .

The first portfolio involves long holding of a European call, a cash

amount of D +XB(τ ) and short selling of one unit of the asset. The
second portfolio contains only one European put. The cash amount
D in the first portfolio is used to compensate the dividends paid to
the lender of the asset due to the short position of the asset. At
expiry, both portfolios are worth max(X − ST , 0). Since both options
are European, they cannot be exercised prior to expiry. By virtue
of law of one price, both portfolios have the same value throughout
the life of the options.

141
Impact of dividends on the lower bound on a European call value
and the early exercise policy of an American call option

Recall the put-call parity relation: c = S − XB(τ ) − D + p. Since

p ≥ 0 and c ≥ 0, we deduce that

c(S, τ ; X, D) ≥ max(S − XB(τ ) − D, 0).

• When S is sufficiently large, the European call almost behaves

like a forward whose value is S − D − XB(τ ). Using the put-call
parity relation: c = p + S − D − XB(τ ), we obtain

p(S, τ ; X, D) → 0 as S → ∞.

142
• Since the call price is lowered due to the dividends of the under-
lying asset, it may be possible that the call price becomes less
than the intrinsic value S − X when the lumped dividend D is
deep enough.

• A necessary condition on D such that c(S, τ ; X, D) may fall below

the intrinsic value S − X is given by

S − X > S − XB(τ ) − D or D > X[1 − B(τ )].

As a necessary condition, suppose D does not satisfy the above
condition (that is, the dividends are not sufficiently deep), then
it is never optimal to exercise the dividend-paying American call
prematurely. This is because C(S, τ ; X, D) ≥ c(S, τ ; X, D) > S −
XB(τ ) − D > S − X when D < X[1 − B(τ )].

143
When D > X[1 − B(τ )], the lower bound S − D − XB(τ ) becomes
less than the intrinsic value S − X. As S → ∞, the European call
price curve falls below the intrinsic value line.

144
Bounds on puts

The bounds for American and European puts can be shown to be

P (S, τ ; X, D) ≥ p(S, τ ; X, D) ≥ max(XB(τ ) + D − S, 0).

The second inequality is easily seen by virtue of the put-call parity
relation and observing c ≥ 0 and p ≥ 0.

• When XB(τ ) + D < X ⇔ D < X[1 − B(τ )], the lower bound
XB(τ ) + D − S may become less than the intrinsic value X − S
when the put is sufficiently deep in-the-money (corresponding
to low value for S). It becomes sub-optimal for the holder of an
American put option to continue holding the put option when
the put value falls below the intrinsic value, so the American put
should be exercised prematurely.
• The presence of dividends makes the early exercise of an Amer-
ican put option less likely since the holder loses the future divi-
dends when the asset is sold upon exercising the American put.
When D > X[1 − B(τ )], it is never optimal to exercise an Amer-
ican put.
145
Lower and upper bounds on the difference of the prices of American
call and put options

• The parity relation cannot be applied to American options due

to the embedded early exercise feature.

First, we assume the underlying asset to be non-dividend paying.

From the put-call parity relation, since P > p and C = c, and putting
D = 0, we have
C − P < S − XB(τ ),
giving the upper bound on C − P .

146
 • Consider the following two portfolios: one contains a European
call plus cash of amount X, and the other contains an American
put together with one unit of underlying asset.

If there were no early exercise of the American put prior to maturity,

the terminal value of the first portfolio is always higher than that
of the second portfolio. If the American put is exercised prior to
maturity, the second portfolio’s value becomes X, which is always
less than c + X. The first portfolio dominates over the second
portfolio, so we have
c + X > P + S.
Further, since c = C when the asset does not pay dividends, the
lower bound on C − P is given by

S − X < C − P.

Remark This result is sharper than the earlier result: C > S − X.

147
Combining the two bounds, the difference of the American call and
put option values on a non-dividend paying asset is bounded by

S − X < C − P < S − XB(τ ).

How to modify the bounds when the underlying asset pays dividend-
s?

• The right side inequality: C − P < S − XB(τ ) also holds for

options on a dividend paying asset since dividends decrease call
value and increase put value.

• The left side inequality has to be modified as: S −D −X < C −P .

• Combining the results, the difference of the American call and

put option values on a dividend paying asset is bounded by

S − D − X < C − P < S − XB(τ ).

148
1.4 American options: Optimal early exercise policies

Non-dividend paying asset

• At any moment when an American call is exercised, its value

immediately becomes max(S − X, 0). The exercise value is less
than max(S − XB(τ ), 0), the lower bound of the call value if
the American call remains alive. The act of exercising prior to
expiry causes a decline in value of the American call.

• Since the early exercise privilege is forfeited, the American and

European call values should be the same when the underlying
asset does not pay any dividend within the life of the American
call option.

• For an American put, it may become optimal to exercise pre-

maturely when S falls to sufficiently low value. Since the gain
in time value of strike is zero when the interest rate is zero, an
American put is never exercised prematurely under zero interest
rate.
149
Dividend paying asset

• When the underlying asset pays dividends, the early exercise of

an American call prior to expiry may become optimal when (i)
S is very high and (ii) the dividends are sizable. Under these
circumstances, it then becomes more attractive for the investor
to acquire the asset rather than holding the option.

– When S is high, the chance of regret of early exercise is low;

equivalently, the insurance value of holding the call is lower.
– When the dividends are sizable, it is more attractive to hold
the asset directly instead of holding the call.
– A necessary condition for optimal early exercise of an Amer-
ican call is D > X[1 − B(τ )]. Note that when D = 0, this
necessary condition cannot be satisfied. To be sufficient for
optimal early exercise, the asset price has to be sufficiently
high.

150
• For an American put, when D is sufficiently high, it may become
non-optimal to exercise prematurely even at very low value of
S (even when the put is very deep-in-the-money). The gain in
the time value of the strike cannot offset the losses in insurance
value of holding the put and the dividends received through
holding the asset.

– A necessary condition for optimal early exercise of an Ameri-

can put is D < X[1−B(τ )]. Note that this necessary condition
is satisfied automatically when D = 0. To be sufficient for
optimal early exercise, the asset price has to be sufficiently
low.

151
 American call American put
time value of strike, loss gain
X
dividend, Dtotal gain loss
insurance value as- loss loss
sociated with hold-
ing of the option
Necessary condition Dtotal > X[1−B(τ )] Dtotal < X[1−B(τ )]
for early exercise
Sufficiently deep-in- Sufficiently high as- Sufficiently low as-
the-money set price set price

(1) when D = 0, never early exercise an American call.

(2) when r = 0, never early exercise an American put.

152
American call on an asset paying discrete dividends

• Since the holder of an American call on an asset paying discrete

dividends will not receive any dividend in between the dividend
times, so within these periods, it is never optimal to exercise
the American call.

• It may be optimal to exercise the American call immediately be-

fore the asset goes ex-dividend. What are the necessary and
sufficient conditions on the size of the dividend and the asset
price level right before the dividend time for optimal early exer-
cise?

153
One-dividend model – Amount of D is paid out at td

• If the American call is exercised at t−

d , the call value becomes
Sd− − X. The asset price drops to Sd+ = Sd− − D right after the
dividend payout.

• It behaves like an ordinary European option for t > t+

d . This is
because when there is no further dividend, it becomes always
non-optimal to exercise the American call.

• The lower bound of the one-dividend American call value at t+

d
is the same lower bound for a European call, which is given by
+ −r(T −t+
d ) , where T − t
+
Sd − Xe d is the time to expiry.

• By virtue of the continuity of the call value across the dividend

date, the lower bound B for the call value at time t−
d should also
be equal to B = Sd+ − Xe−r(T −td) = (Sd− − D) − Xe−r(T −td).

154
By comparing the call value (continuation) with the early exercise
proceed: Sd− − X, we deduce

(i) If Sd− − X ≤ B,

Sd− − X ≤ (Sd− − D) − Xe−r(T −td)

or D ≤ X[1 − e−r(T −td)]
it is never optimal to exercise since exercising leads to a drop in
call value.

(ii) When the discrete dividend D is sufficiently sizable such that

D > X[1 − e−r(T −td)],

it may become optimal to exercise at t−d when the asset price
Sd− is above some threshold value Sd∗.

In summary, a necessary condition for early exercise of the American

call is that the dividend amount D is larger than the loss of the value
of the strike price X.
155
Let Cd(Sd−, T − t−
d ) denote the American call price at t −
d , where time
to expiry is T − t−
d . Suppose the American call stays alive across td ,
by continuity of call value across the dividend date, we have
Cd(Sd−, T − t−
d ) = c(Sd
+
, T − t +
d ) = c(Sd
−
− D, T − t +
d ).
The optimal exercise price Sd∗ is the solution to

c(Sd− − D, T − t+
d ) = S −
d − X.
When D > X[1 − e−r(T −td)], then the American call price function
right before td consists of two segments:
{
c(Sd− − D, T − t+ ) when S − ∗
d < Sd .
Cd(Sd−, T − t−
d)=
d
Sd− − X when Sd− ≥ Sd∗
The American call price at t < td can be obtained by the risk neutral
valuation formula with the known value at td. Thus, Sd∗ depends
on D, which decreases when D increases. This is because the price
curve of c(Sd− − D, T − t+
d ) is lowered and it cuts the intrinsic value
line ℓ1 at a lower value of Sd∗. Financially, the holder chooses to
receive the asset even at a lower asset value when the dividend is
deeper.
156
Determination of Sd∗ (potential early exercise at t−
d when D is suffi-
ciently deep)
c ( S d−− D, t d+ ) l1 l2

S d−
X S d∗ Xe −r(T−t d)
+D

The European call price function V = c(S –d − D, t +

d ) falls below the
exercise payoff line ℓ1 : E = S d– − X. Note that ℓ1 lies to the left
of the lower bound value line ℓ2 : B = S – −r(T −td ) when
d − D − Xe
D > X[1 − e−r(T −td)]. Here, S ∗d is the value of S − d at which the
European call price curve cuts the exercise payoff line ℓ1.

157
Summary of early exercise policies of American calls

• With no dividend, the decision of early exercise of an American

option (call or put) depends on the competition between the
time value of X and the loss of insurance value associated with
the holding of the option.
• Early exercise of non-dividend paying American call is non-optimal
since this leads to the loss of insurance value of the call plus the
loss of time value of X.
• For an American call on a discrete dividend paying asset, it may
become optimal to exercise at time right before the ex-dividend
time, provided that the dividend amount is sizable and the call
is sufficiently deep in-the-money (Sd− > Sd∗). The critical asset
price Sd∗ is a decreasing function of the size of dividend. Early
exercise of the American call at a lower asset price level leads
to a greater loss of insurance value but the loss is offset by the
more sizable dividend received.

158
Continuous dividend model

Under constant dividend yield q, the dividend amount received during

(t, t + dt) from holding one unit of asset is qSt dt. Also, e−q(T −t) unit
of the asset at time t will become one unit at time T through the
accumulation of the dividends into purchase of the asset.

Why do we consider dividend yield model?

• It is considered as a continuous approximation to the discrete

dividends model. Otherwise, pricing under the discrete n-dividend
model requires the joint distribution of asset prices on all divi-
dend dates: Std , Std , · · · , Stdn .
1 2

• The foreign money market account, which serves as the under-

lying asset in exchange options, earns the foreign interest rate
rf that can be visualized as the dividend yield.

159
When the underlying asset pays dividend yield
( q, the lower bound
) of
a European call c(S, τ ; X, q) becomes max Se−qτ − Xe−rτ , 0 . As S
becomes sufficiently high, Se−qτ − Xe−rτ becomes less than S − X.
Actually, as S → ∞, the value of the European call tends to that of
the forward, where

c(S, τ ; X, q) → Se−qτ − Xe−rτ .

160
Smooth pasting condition at S ∗(τ ) under continuous dividend yield
model

Value matching at S ∗(τ ) : C(S ∗(τ ), τ ) = S ∗(τ ) − X. (i)

∂C ∗
Smooth pasting at S ∗(τ ) : (S (τ ), τ ) = 1. (ii)
∂S
S ∗(τ ) can be visualized as the lowest asset price at which the Amer-
ican call does not depend on the time to expiry. That is,
∂C
= 0 at S = S ∗(τ ). (iii)
∂τ
Find the total derivative of the American call price function with
respect to τ at (S ∗(τ ), τ ) satisfying the value matching condition:
d [ ∗ ] ∂C ∗ ∂C ∗ dS ∗(τ ) dS ∗(τ )
C(S (τ ), τ ) = (S (τ ), τ ) + (S (τ ), τ ) =
dτ ∂τ ∂S dτ dτ
∂C ∗
By eq(iii), we obtain (S (τ ), τ ) = 1. The smooth pasting condi-
∂S
tion (ii) can also be derived from optimality of early exercise (to be
visualized as the first order derivative condition).
161
American call on a continuous dividend yield paying asset

The option price curve of a longer-lived American call will be above

that of its shorter-lived counterpart for all values of S. The upper
price curve cuts the intrinsic value tangentially at a higher critical
asset value S ∗(τ ). Hence, S ∗(τ ) for an American call is an increasing
function of τ .

162
Properties of the optimal early exercise boundary S ∗(τ ) of an Amer-
ican call under continuous dividend yield model

X = strike price, r = riskfree interest rate, q = constant dividend

yield, σ = volatility of asset price
( )
r µ+
S ∗(0+) = X max 1, ∗
, S∞ = X,
q µ+ − 1
( ) √ ( )2
− r−q− 2 σ2 + σ2
r−q− 2 + 2σ 2r
0 < µ+ =
σ2

163
Stopping region = {(S, τ ) : S ≥ S ∗(τ )}, inside which the American
call should be optimally exercised. When S < S ∗(τ ), it is optimal
for the holder to continue holding the American call option.

1. At a higher value of τ , critical asset price S ∗(τ ) is larger since the

American call option has to be deeper in-the-money to induce
optimal early exercise. S ∗(τ ) is monotonically increasing with
respect to τ with
( )
r µ+
S ∗(0+) = X max 1, ∗ =
and S∞ X. The determination
q µ+ − 1
∗ requires a pricing model of the perpetual American call
of S∞
option.

2. S ∗(τ ) is a continuous function of τ when the asset price process

is continuous.

3. S ∗(τ ) ≥ X for τ ≥ 0
Suppose S ∗(τ ) < X, then the early exercise proceed S ∗(τ ) − X
becomes negative. This must be ruled out.
164
One-dividend paying model for an American put

• A single dividend D is paid at td.

• Never exercise immediately prior to the dividend payment for

time t < td, since gain in the time value of X is X[er(td−t) − 1]
cannot offset the loss in dividend. We would like to find ts such
that the holder of the American put is indifferent to the time
value of strike and the dividend. That is,
( )
D
ln 1 + X
X[er(td−ts) − 1] = D giving ts = td − .
r
• When t < ts, early exercise is optimal when S falls below some
critical asset price S ∗(t).

165
The behavior of the optimal exercise boundary S ∗(t) as a function
of t for a one-dividend American put option. Note that S ∗(T ) = X
for an American put option whose underlying asset becomes non-
dividend paying after td.

166
In summary, the optimal exercise boundary S ∗(t) of the one-dividend
American put model exhibits the following behavior.

(i) When t < ts, S ∗(t) first increases then decreases smoothly with
increasing t until it drops to the zero value at ts. When the
calendar time is well before ts, S ∗(t) is increasing as it follows
a general trend of increasing monotonically in time. However,
when the calendar time comes closer to ts, S ∗(t) decreases in
time in order to adopt the trend that S ∗(t) falls to zero when t
reaches ts.

(ii) S ∗(t) stays at the zero value in the time interval [ts, td].

(iii) When t ∈ (td, T ], the American put behaves like the European
put counterpart since there will be no more dividend. As a result,
S ∗(t) is a monotonically increasing function of t with S ∗(T ) = X.

167
Here, td < t1 < t2 < T . The put price curve at time t1 intersects
the intrinsic value line tangentially at S ∗(t1). We observe: S ∗(t1) <
S ∗(t2) < X. At longer time to expiry, the American put has to be
deeper in-the-money (lower asset price) in order to induce optimal
early exercise.

168