Formulas for MFE

1 Chapter 9 Parity and Other Relationships

1.1 Options on Stock

C(K , T ) = P (K , T ) + [So − PV0,T (Div)] − e−r TK

C(K , T ) = P (K , T ) + Soe−δ T − PV0,T (K)

1.2 Options on Currencies

From Chapter 5, dollar forward price for a euro is F0,T = x0e(r −re u r o )T , where x0 is the current
exchange rate denominated as $/euro.

C(K , T ) − P (K , T ) = x0e−re u r o T − Ke−rT

1.3 Options on Bonds

C(K , T ) = P (K , T ) + [B0 − PV0,T (Coupons)] − PV0,T (K)

1.4 Generalized Parity and Exchange Options

C(ST , QT , 0) = max (0, ST − QT )

P (ST , QT , 0) = max (0, QT − St)
P P
C(St , Qt , T − t) − P (St , Qt , T − t) = Ft,T (S) − Ft,T (Q)

1.5 Currency Options

 
1 1
C$ (x0, K , T ) = x0KP f , ,T
x0 K

1.6 Maximum and Minimum Options Prices

S > CAmer (S , K , T ) > CEur (S , K , T ) > max [0, PV0,T (F0,T ) − PV0,T (K)]
K > PAm er (S , K , T ) > PEur (S , K , T ) > max [0, PV0,T (K) − PV0,T (F0,T )]

1.7 Early Exercise

Exercise on a call is not optimal when:

K − PVt,T (K) > PVt,T (Div)

1
2 Section 2

1.8 Different Strike Prices

K1 < K2 < K3
1. C(K1) > C(K2)
2. P (K2) > P (K1)
3. C(K1) − C(K2) 6 K2 − K1
4. P (K2) − P (K1) 6 K2 − K1
C(K1) − C(K2) C(K2) − C(K3)
5. K2 − K1
> K3 − K2
P (K2) − P (K1) P (K3) − P (K2)
6. K2 − K1
6 K3 − K2

2 Binomial Option Pricing: I

2.1 The Binomial Solution

Cu − Cd uCd − d Cu
∆ = e−δh B = e−rh
S(u − d) u−d

e(r −δ)h − d u − e(r −δ)h
∆S + B = e−rh Cu + Cd
u−d u−d
(r −δ)h
No arbitrage ⇒ u > e >d

2.2 Risk-Neutral Pricing

e(r −δ)h − d
p∗ =
u−d
C = e−r h[p∗Cu + (1 − p∗)Cd]
√
u = e(r −δ)h+σ√h
d = e(r −δ)h−σ h

2.3 Options on Currencies

√
ux = xe(r −rf )h+σ√h
dx = xe(r −rf )h−σ h
(r −r f )h
e −d
p∗ =
u−d

2.4 Options on Futures Contracts

√
u = eσ√ h
d = eσ h
1−d
p∗ =
u−d
The Black-Scholes Formula 3

3 Binomial Option Pricing: II

√ σ
σh = σ h Example: σmonthly = √
12

4 The Black-Scholes Formula

4.1 Calls and Puts

C(S , K , σ, r, T , δ) = Se−δTN (d1) − Ke−rTN (d2)

1
ln(S/K) + (r − δ + 2 σ 2)T
d1 = √
σ T
√
d2 = d1 − σ T

P (S , K , σ, r, T , δ) = Ke−r TN ( − d2) − Se−δTN ( − d1)

4.2 Options on Stocks with Discrete Dividends

P
F0,T (S) = S0 − PV0,T (Div)

4.3 Options on Currencies

P
F0,T (S) = x0e−rfT

C(x, K , σ, r, T , r f ) = xe−rfTN (d1) − Ke−r TN (d2)

1
ln(x/K) + (r − r f + 2 σ 2)T
d1 = √
σ T
P (x, K , σ, r, T , r f ) = C(x, K , σ, r, T , r f ) + Ke−r T − xe−rfT

4.4 Options on Futures

δ =r
This is know as the Black Formula.

4.5 Greek Measures for Portfolios

n
X
∆p ortfolio = ωi∆i
i=1
Holds true for other greeks as well.
4 Section 6

4.6 Option Elasticity

ǫ is the change in the stocks price. Ω is option elasticity.

ǫ∆
% change in option price C S∆
Ω≡ = ǫ =
% change in sto ck price
S
C

4.7 Volatility of an option

σoption = σsto ck × |Ω|

4.8 Risk Premium of an Option

γ is the expected return of the option.

γ − r = (α − r) × Ω

4.9 Calendar Spreads

Calendar spreads: buy and sell options with different expirations.

5 The Standard Normal Distribution

1 − 1 x2
Z x
φ(x) ≡ √ e 2 N (x) ≡ φ(x)dx
2π −∞

6 Option Greeks

Important identities: N (x) = 1 − N ( − x) Se−δTN ′(d1) = Ke−rT N ′(d2)

6.1 Delta (∆)

∂C(S , K , σ, r, T − t, δ)
∆call = = e−δ(T −t)N (d1)
∂S

∂P (S , K , σ, r, T − t, δ)
∆put = = e−δ(T −t)N ( − d1)
∂S

6.2 Gamma (Γ)

Market-Making and Delta-Hedging 5

∂ 2C(S , K , σ, r, T − t, δ) e−δ(T −t)N ′(d1)

Γcall = Γput = = √
∂ 2S Sσ T − t

6.3 Theta (θ)

∂C(S , K , σ, r, T − t, δ) Ke−r(T −t)N ′(d2)σ

θcall = = δSe−δ(T −t)N (d1) − rKe−r(T −t)N (d2) − √
∂t 2 T −t
∂P (S, K , σ, r, T − t, δ)
θput = = θcall + rKe−r(T −t) − δSe−δ(T −t)
∂t

6.4 Vega

∂C(S , K , σ, r, T − t, δ) √
Vegacall = Vegaput = = Se−δ(T −t)N ′(d1) T − t
∂σ

6.5 Rho (ρ)

∂C(S , K , σ, r, T − t, δ)
ρcall = = (T − t)Ke−r(T −t)N (d2)
∂r
∂P (S , K , σ, r, T − t, δ)
ρput = = (T − t)Ke−r(T −t)N ( − d2)
∂r

6.6 Psi (ψ)

∂C(S , K , σ, r, T − t, δ)
ψcall = = − (T − t)Se−δ(T −t)N (d1)
∂δ
∂P (S , K , σ, r, T − t, δ)
ψput = = (T − t)Se−δ(T −t)N ( − d1)
∂δ

7 Market-Making and Delta-Hedging

7.1 Understanding Market-Makers Profit

1
∆t(St+h − St) − [∆t(St+h − St) + 2 (St+h − St)2Γt + θh] − rh[∆tSt − C(St)] =
 
1
− 2 ǫ2Γt + θth + rh[∆tSt − C(St)]

One stardard deviation move: ǫ2 = σ 2St2h. In that case we have

 
1
Market-maker profit= − 2 σ 2St2Γt + θ + r[∆tSt − C(St)] h. Set profit equal to zero and rearrange
terms to get:
6 Section 8

1 2 2
2
σ St Γt + rSt∆t + θ = rC(St)

7.2 Re-hedgeing

Re-hedge every h, market has moved xi standard deviations.

1
Rh,i = 2 S 2σ 2Γ(x2i − 1)h
1
Var(Rh,i) = 2 (S 2σ 2Γh)2 Example Variance for a day for a daily re-hedger:
1
Var(R1/365,1) = 2 (S 2σ 2Γ/365)2

7.3 Greeks in the Binomial Model

Cu − Cd
∆(S , 0) = e−δh
uS − dS
∆(uS , h) − ∆(dS , h)
Γ(Sh , h) =
uS − dS
ǫ = udS − S
1
C(udS , 2h) − ǫ∆(S , 0) − 2 ǫ2Γ(S , 0) − C(S , 0)
θ(S , 0) =
2h

8 Exotic Options: I

8.1 Asian Options

max [0, ± (G(T ) − K)] where G(T ) is some sort of average and the sign depends on it being a call or
put.

8.2 Barrier Options

1. Knock-out. Go out of existence when a barrier is crossed.

2. Knock-in. Go into existence when a barrier is crossed.
3. Rebate. Make a fixed payment if barrier is crossed.

“Knock-in” option + “Knock-out” option = Oridinary Option

8.3 Compound Options

max [C(St , K , T − t1) − x, 0].

Compound Option Parity:

The Lognormal Distribution 7

CallOnCall(S , K , x, σ, r, t1, t2, δ) − PutOnCall(S , K , x, σ, r, t1, t2, δ) + xe−r t1 = BSCall(S , K , σ, r, t2, δ)

8.4 Gap Options

K1 strike. K2 trigger.

C(S , K1, K2, σ, r, T , δ) = Se−δTN (d1) − K1e−r TN (d2)

1
ln(Se−δT /K2e−rT ) + 2 σ 2T
d1 = √
σ T
√
d2 = d1 − σ T

8.5 Exchange Options

max (0, ST − Kt)

C(S , K , σ, r, T , δ) = Se−δSTN (d1) − K1e−δK TN (d2)

1
ln(Se−δST /Ke−δK T ) + 2 σ 2T
d1 = √
σ T
√
d2 = d1 − σ T
p
σ= σS2 + σK
2
− 2ρσSσK

9 The Lognormal Distribution

9.1 The Normal Distribution

 2
1 x− µ
1 −
φ(x; µ, σ) ≡ √ e 2 σ
σ 2π

9.2 Sum of Normal Random Variables

xi ∼ N (µi , σi2) and Cov(xi , x j ) = σi, j . σij = ρijσiσ j

n n
!
X X
E ωi x i = ωiµi
i=1 i=1

n n X
n
!
X X
Var ωixi = ωiω jσij
i=1 i=1 j =1
8 Section 10

9.3 The Lognormal Distribution

Continously compounded return definition:

R(0, t) = ln(St/S0) or St = S0eR(0,t)

1
m+ 2 v 2 2 2
If x ∼ N (m, v 2) then E(ex) = e and Var(ex) = e2m+v (ev − 1)

ln(St/S0) ∼ N [(α − δ − 0.5σ 2)t, σ 2t]

√ 2
√
ln(St/S0) = (α − δ − 0.5σ 2)t + σ t z St = S0e(α−δ −0.5σ )t+σ t z

E(St) = S0e(α−δ)t

9.4 Lognormal Probablity Calculations

Prob(St < K) = N ( − dˆ2) Prob(St > K) = N (dˆ2) where dˆ2 is the standard Black-Schoes argument
with r → α.

N ( − dˆ1) N (dˆ1)
N (St |St < K) = Se(α−δ)t N (St |St > K) = Se(α−δ)t
N ( − dˆ2) N (dˆ2)

10 Monte Carlo Valuation

10.1 Using Sums of Uniformly Distributed Random Variables

12
X
Z̃ = ui − 6
i=1

10.2 Monte Carlo Valuation

√ P 1 √ 1 Pn
1
(α−δ − 2 σ 2)T +σ h [ n (α−δ − 2 σ 2)T +σ T [ √ i=1Z(i)]
i=1Z(i)]
ST = S0e =e n

n
1
X
V (S0, 0) = n e−r T V (STi , T )
i=1

10.3 Control Variate Method

A∗ = Ā + β(G − Ḡ )

where β = Cov(Ā , Ḡ )/Var(Ḡ )

Brownian Motion and Itô’s Lemma 9

11 Brownian Motion and Itô’s Lemma

11.1 Black-Scholes Assumptions about Stock Prices

dS(t)
S(t)
= αdt + σdZ(t)

ln[S(T )] ∼ N (ln[S(0)] + [α − 0.5σ 2]T , σ 2T )

11.2 Brownian Motion

• Z(0) = 0
• Z(t + s) − Z(t) ∼ N (0, s)
• Z(t + s1) − Z(t) is independant of Z(t) − Z(t − s2) s1, s2 > 0
• Z(t) is continuous

Above implies that Z(t) is a martingale (i.e. E[Z(t + s)|Z(t)] = Z(t)

small h, Y (t) = { − 1, 1}, E[Y (t)] = 0, Var[Y (t)] = 1,

√
Z(t + h) − Z(t) = Y (t + h) h

h = T /n
n
" #
√ 1 X
Z(T ) − Z(0) = T √ Y (ih)
n i=1

√
dZ(t) = Y (t) dt

n
" #
√ 1 X
Z T
Z(T ) = Z(0) + lim T √ Y (ih) → Z(0) + dZ(t)
n→∞ n i=1 0

11.3 Properties of Brownian Motion

n n  n
X X √ 2 X
2
lim (Z[ih] − Z[(i − 1)/h])2 = lim h Yih = lim hYih =T
n→∞ n→∞ n→∞
i=1 i=1 i=1

Thus it has finite quadratic variation so:

n
X
lim (Z[ih] − Z[(i − 1)/h])n = 0 for n > 2.
n→∞
i=1
10 Section 11

But infinite total variation:

n
X
lim |Z[ih] − Z[(i − 1)/h]| = ∞
n→∞
i=1

11.4 Arithmetic Brownian Motion

X(T ) − X(0) = αT + σZ(T )

dX(t) = αdt + σdZ(t)

X(T ) − X(0) ∼ N (αT , σ 2T )

11.5 The Ornstein-Uhenbeck Process

dX(t) = λ[α − X(t)]dt + σdZ(t)

11.6 Geometric Brownian Motion

dX(t) = α[X(t)]dt + σ[X(t)]dZ(t)

dX(t)
= αdt + σdZ(t)
X(t)

ln[X(t)] ∼ N (ln[X(0)] + (α − 0.5σ 2)t, σ 2t)

2
√
X(t) = X(0)e(α−0.5σ )t+σ t Z

E[X(t)] = X(0)eαt

11.7 Multiplication Rules

dt × dZ = 0

(dt)2 = 0

(dZ)2 = dt

dZ × dZ ′ = ρdt

11.8 The Sharpe Ratio

αi − r
Sharpe ratioi =
σi
The Black-Scholes Equation 11

11.9 The Risk Neutral Process

dS(t)
= (α − δ)dt + σdZ(t)
S(t)

Z̃ (t) generates a martingale in utility terms for a risk-averse investor.

dS(t)
= (r − δ)dt + σd Z̃ (t)
S(t)

dZ̃ (t) = dZ(t) + ηdt where η = (α − r)/σ

11.10 Itô’s Lemma

n o
dS(t) = α̂[S(t), t] − δˆ[S(t), t] dt + σ̂ [S(t), t]dZ(t)

1
dC(S , t) = CSdS + 2 CS S (dS)2 + Ctdt
n o
1
dC(S , t) = [α̂(S , t) − δˆ(S , t)]CS + 2 σ̂ (S , t)2CS S + Ct dt + σ̂ (S , t)CSdZ

11.11 Valuing a Claim on S a

1
P [a(r −δ)+ 2 a(a−1)σ 2]T
F0,T [S(T )a] = e−rTS(0)ae
1
[a(r −δ)+ 2 a(a−1)σ 2]T
F0,T [S(T )a] = S(0)ae

12 The Black-Scholes Equation

dS
= (α − δ)dt + σdZ
S

Option V [S(t), t]. Invest W in bonds that pay return r.

dW = rWdt

Total investment in option, stocks (N shares) and bonds should be zero.

I = V (S , t) + NS + W = 0
1
dI = dV + N (dS + δSdt) + dW = Vtdt + VSdS + 2 σ 2S 2VSSdt + N (dS + δSdt) + rWdt

Delta-hedge so N = − VS . Bonds: W = VSS − V . So,

1
dI = Vt + 2 σ 2S 2VS Sdt − VSδSdt + r(VSS − V )dt
12 Section 13

Zero-investment, zero-risk portfolio so dI = 0.

1
Vt + 2 σ 2S 2VS S + (r − δ)SVS − rV = 0

12.1 Risk Neutral Pricing

dS
= (r − δ)dt + σdZ̃
S

1 1
E(dV ) = Vt + σ 2S 2VSS + (α − δ)SVS
dt 2

E ∗(dS) = (r − δ)dt
1 1 1
dt
E ∗(dV ) = Vt + 2 σ 2S 2VS S + (r − δ)SVS so dt
E ∗(dV ) = rV

13 Exotic Options: II

13.1 All-Or-Nothing Options

CashCall(S , K , σ, r, T − t, δ) = e−r(T −t)N (d2)

CashPut(S , K , σ, r, T − t, δ) = e−r(T −t)N ( − d2)

AssetCall(S , K , σ, r, T − t, δ) = e−δ(T −t)SN (d1)

AssetPut(S , K , σ, r, T − t, δ) = e−δ(T −t)SN ( − d1)

13.2 Ordinary Options and Gap Options

BSCall(S , K , σ, r, T − t, δ) = AssetCall(S , K , σ, r, T − t, δ) − K × CashCall(S , K , σ, r, T − t, δ)

BSPut(S , K , σ, r, T − t, δ) = K × CashPut(S , K , σ, r, T − t, δ) − AssetPut(S , K , σ, r, T − t, δ)

Gap option that pays S − K1 if S > K2.

AssetCall(S , K2, σ, r, T − t, δ) − K1 × CashCall(S , K2, σ, r, T − t, δ)

Interest Rate Models 13

14 Volatility

dSt/St = (α − δ)dt + σ(St , Xt , t)dZ

ǫt+h = ln(St+h/St)

n
2 1
σˆH
X
= ǫ2i
(n − 1)
i=1

14.1 ARCH

ln(St/St−h) = (α − δ − 0.5σ 2)h + ǫt

Var(ǫt) = σ 2h

15 Interest Rate Models

15.1 Behavior of Bonds and Interest Rates

dP
= α(r, t)dt + q(r, t)dZ
P

dr = a(r)dt + σ(r)dZ

15.2 Impossible Bond Pricing Model

P (t, T ) = e−r(T −t)

15.3 An Equilibrium Equation for Bonds

1 ∂ 2P 1 ∂ 2P
 
∂P 2 ∂P ∂P 2 ∂P ∂P
dP (r, t, T ) = dr + (dr) + dt = a(r) dr + σ(r) + dt + σ(r)dZ
∂r 2 ∂r 2 ∂t ∂r 2 ∂r 2 ∂t ∂r

1 ∂ 2P
 
1 ∂P 2 ∂P
α(r, t, T ) = a(r) dr + σ(r) +
P (r, t, T ) ∂r 2 ∂r 2 ∂t

1 ∂P
q(r, t, T ) = − σ(r)
P (r, t, T ) ∂r
14 Section 16

dP (r, t, T )
= α(r, t, T )dt − q(r, t, T )dZ
P (r, t, T )

Delta-hedged portfolio

dI = N [α(r, t, T1)dt − q(r, t, T1)dZ]P (r, t, T1) + [α(r, t, T2)dt − q(r, t, T2)dZ]P (r, t, T2) + rWdt

P (r, t, T2) q(r, t, T2) P (r, t, T2)

Set N = − =− r and dI = 0
P (r, t, T1) q(r, t, T1) Pr(r, t, T2)

Thus Sharpe ratio for the two bonds is equal

α(r, t, T1) − r α(r, t, T2) − r

=
q(r, t, T1) q(r, t, T2)

α(r, t, T ) − r
φ(r, t) =
q(r, t, T )

1 ∂ 2P ∂P ∂P
σ(r)2 2 + [a(r) + σ(r)φ(r, t)] + − rP = 0
2 dr ∂r ∂t

The risk-neutral process for the interest rate:

dr = [a(r) + σ(r)φ(r, t)]dt + σ(r)dZ

P [t, T , r(t)] = Et∗[e−R(t,T )]

Z T
R(t, T ) = r(s)ds
t

15.4 Delta-Gamma Approximations for Bonds

1 ∗
E (dP ) = rP
dt

16 Equilibrium Short-Rate Bond Price Models

16.1 The Rendelman-Bartter Model

Equilibrium Short-Rate Bond Price Models 15

dr = adt + σdZ

16.2 The Vasicek Model

dr = a(b − r)dt + σdz

1 2 ∂ 2P ∂P ∂P
σ + [a(b − r) − σφ] + − rP = 0
2 ∂r 2 ∂r ∂t

P [t, T , r(t)] = A(t, T )e−B(t,T )r(t)

2 2
A(t, T ) = er̄(B(t,T )+t−T )−B σ /4a

B(t, T ) = (1 − e−a(T −t))/a

r̄ = b + σφ/a − 0.5σ 2/a2

r̄ is the yield to maturity for an infinite length bond.

16.3 The Cox-Ingersoll-Ross Model

√
dr = a(b − r)dt + σ r dz
√
Sharpe Ratio: φ(r, t) = φ̄ r /σ

1 2 ∂ 2P ∂P ∂P
σ + [a(b − r) − rφ̄ ] + − rP = 0
2 ∂r 2 ∂r ∂t

P [t, T , r(t)] = A(t, T )e−B(t,T )r(t)

" #
2γe(a− φ̄+ γ)(T −t)/2
A(t, T ) =
(a − φ̄ + γ)(e γ(T −t) − 1) + 2γ

2(e γ(T −t) − 1)

B(t, T ) =
(a − φ̄ + γ)(eγ(T −t) − 1) + 2γ
p
γ= (a − φ̄ )2 + 2σ 2

16.4 Bond Options, Caps, and The Black Model

• Pt(T , T + s) is zero-coupon bond price at time t purchased at time T and paying $1 at time T + s
16 Section 16

• If t = T then P (T , T + s) is the spot price

• If t < T then Pt(T , T + s) is a forward price Ft,T [(P (T , T + s)]

Call option payoff = max [0, P (T , T + s) − K]

Ft,T [P (T , T + s)] = P (t, T + s)/P (t, T )

Volatility = Var(ln(Ft,T [P (T , T + s)]))

C[F , P (0, T ), σ, T ] = P (0, T )[FN (d1) − KN (d2)]

ln(F /K) + 0.5σ 2T

d1 = √
σ T
√
d2 = d1 − σ T

where F = F0,T [P (T , T + s)]

P (0, T )
R0(T , T + s) = −1
P (0, T + s)

Foward rate agreement (FRA)

Payoff to FRA = RT (T , T + s) − R0(T , T + s)

Call option on FRA is a caplet.

Payoff to caplet = max [0, RT (T , T + s) − KR]

If settled at time T , the option pays:

1
max [0, RT (T , T + s) − KR]
1 + RT (T , T + s)

Let RT = RT (T , T + s)

   
R T − KR 1 1
(1 + KR)max 0, = (1 + KR)max 0, −
(1 + KT )(1 + KR) 1 + KR 1 + R T

cap is a collection of caplets

Jensen’s Inequality 17

Cap payment at time ti+1 = max [0, Rti(ti , ti+1) − KR]

16.5 A Binomial Interest Rate Model

Pi(i, i + 1; j) = e−ri(i,i+1; j)h

P0(0, 1; 0) = e−r h

P0(0, 2; 0) = e−r h[pe−ruh + (1 − p)e−rdh] = e−r h[pP1(1, 2; 1) + (1 − p)P1(1, 2; 0)]

 Pn 
Using risk neutral E ∗ e− i=0rih

Yields= − ln(P (0, T ))/T

16.6 The Black-Derman-Toy Model

√
Aeσ h
Distance between up node and down node is √
Ae −σ h

Yield: y[h, T , r(h)] = P [h, T , r(h)]−1/(T −h) − 1

y(h, T , ru) √
 
Yield volatility = 0.5 × ln / h
y(h, T , rd)

P (0, 1) × (0.5 × P (1, 2; Ru) + 0.5 × P (1, 2; Rd)) = P (0, 2)

17 Interest Rates

Effective annual rate: r in (1 + r)n

Continuously compounded rate: r in er n

18 Jensen’s Inequality

If f (x) is convex: E[f (x)] > f [E(x)]

If f (x) is concave: E[f (x)] 6 f [E(x)]