Mathematical Methods for Financial Markets

by M. Jeanblanc, M. Yor and M. Chesney

Springer Verlag 2009, Chapter 2.

Definition 1.4.1.1
The continuous process X is said to be a standard Brownian motion, if:

1. The process X has stationary and independent increments

2. For any t > 0, the r.v. Xt follows the N (0, t) law
3. X0 = 0

2.3.1 The Model

The Black and Scholes model assumes that there is a riskless asset with interest
rate r and that the dynamics of the price of the underlying asset are
dSt = St (µdt + σdBt )
under the historical probability P. Here, the risk-free rate r, the trend µ and the
volatility σ are supposed to be constant (note that, for valuation purposes, µ may
be an F-adapted process). In other words, the value at time t of the risky asset is
σ2
St = S0 exp(µt + σBt − t).
2

Indeed, by applying Ito’s lemma to the function ln(·):

µ ¶ µ ¶
1 1 1 2 2 σ2
d ln(St ) = dSt + − 2 σ St dt = µ − dt + σBt
St 2 St 2
Proposition 2.3.1.2
In the Black and Scholes model, there exists a unique e.m.m. Q, precisely Q|Ft =
exp(−θBt − 12 θ2 t)P|Ft where θ = µ−r
σ
is the risk-premium. The risk-neutral dy-
namics of the asset are
dSt = St (rdt + σdWt )
where W is a Q-Brownian motion.
Theorem 2.3.2.1 Black and Scholes formula:
Let dSt = St (µdt + σdBt ) be the dynamics under the historical probability P of
the price of a risky asset and assume that the interest rate is a constant r. The
value at time t of a European call with maturity T and strike K is CE (x, t) where
h ³ x ´i
CE (x, t) = xN d1 ,T − t (1)
Ke−r(T
h ³
−t)
´i
x
− Ke−r(T −t) N d2 −r(T −t)
,T − t
Ke

1
 R di 1 2
where N(di ) = √1 e− 2 y dy for i = 1, 2
−∞ 2π

1 1√ 2 √
d1 (y, u) = √ ln(y) + σ u, σ2u ,
d2 (y, u) = d1 (y, u) −
σ2u 2
√
where we have written σ 2 so that the formula does not depend on the sign of σ.
Proof:
Let (α, β) be a replicating portfolio and

Vt = αt Ct + βt St

We assume that the value of this portfolio satisfies the self-financing condition, i.e.,

dVt = αt dCt + βt dSt

Then, assuming that Ct is a smooth function of the underlying value and of time,
i.e., Ct = C(St , t), by relying on Itô’s lemma the differential of V is obtained:
1
dVt = αt (∂x CdSt + ∂t Cdt + σ 2 St2 ∂xx Cdt) + βt dSt ,
2
where ∂t C (resp. ∂x C ) is the derivative of C with respect to the second variable
(resp. the first variable) and where all the functions C, ∂x C, . . . are evaluated at
(St , t). From αt = (Vt − βt St )/Ct , we obtain

dVt = ((Vt − βt St )(Ct )−1 ∂x C + βt )σSt dBt (2)

µ µ ¶ ¶
Vt − βt St 1 2 2
+ ∂t C + σ St ∂xx C + µSt ∂x C + βt St µ dt.
Ct 2

If this replicating portfolio is risk-free, one has dVt = Vt rdt: the martingale part
on the right-hand side vanishes, which implies

βt = (St ∂x C − Ct )−1 Vt ∂x C

and µ ¶
Vt − βt St 1
∂t C + σ 2 St2 ∂xx C + St µ∂x C + βt St µ = rVt . (3)
Ct 2
Using the fact that αt ∂x C + βt = 0, i.e.

(Vt − βt St )(Ct )−1 ∂x C + βt = 0

we obtain that the term which contains µ, i.e.,

µ ¶
Vt − βSt
µSt ∂x C + βt
Ct

2
vanishes. After simplifications, we obtain
µ ¶µ ¶
St ∂x C 1 2 2
rC = 1+ ∂t C + σ x ∂xx C (4)
C − St ∂x C 2
µ ¶
C 1 2 2
= ∂t C + σ x ∂xx C (5)
C − St ∂x C 2
and therefore the PDE evaluation
1
∂t C(x, t) + rx∂x C(x, t) + σ 2 x2 ∂xx C(x, t)
2
= rC(x, t), x > 0, t ∈ [0, T [ (6)

is obtained. Now,
N(d1 )
βt = Vt ∂x C(S∂x C − Ct )−1 = V0 −rT
. (7)
Ke N(d2 )
Note that the hedging ratio is
βt
= −∂x C(St , t) .
αt
Reading carefully the Black and Scholes (1973) paper, The pricing of options and
corporate liabilities (Journal of Political Economy), it seems that the authors as-
sume that there exists a self-financing strategy (−1, βt ) such that dVt = rVt dt,
which is not true, and the portfolio with value −Ct + St N(d1 ) = Ke−r(T −t) N(d2 )
is not risk-free.
The solution of the PDE (eq. (6)) with terminal condition

C(x, T ) = (x − K)+

is the Black-Scholes formula.

Final comment: Another way to find the solution is to compute the conditional
expectation under the equivalent martingale measure of the discounted terminal
payoff.
In a Black and Scholes model, the price of a European option is given by:

CE (S0 , T ) = EQ (e−rT (ST − K)1{ST ≥K} ) (8)

−rT −rT
= EQ (e ST 1{ST ≥K} ) − e KQ(ST ≥ K). (9)
¡ ¢
Hence, if k = σ1 ln(K/x) − (r − 21 σ 2 )T , using the symmetry of the Gaussian law,
one obtains
³ ³ x ´´
Q(ST ≥ K) = Q(WT ≥ k) = Q(WT ≤ −k) = N d2
Ke−rT
where the function d2 is given in Theorem 2.3.2.1.

3
From the dynamics of S, one can write:
µ ¶
−rT σ2
e EQ (ST 1{ST ≥K} ) = S0 EQ 1{WT ≥k} exp(− T + σWT ) .
2
2
The process (exp(− σ2 t + σWt ), t ≥ 0) is a positive Q-martingale with expectation
equal to 1. Let us define the probability Q∗ by its Radon-Nikodým derivative with
respect to Q:
σ2
Q∗ |Ft = exp(− t + σWt ) Q|Ft .
2
Hence,
e−rT EQ (ST 1{ST ≥K} ) = S0 Q∗ (WT ≥ k) .
ct = Wt −σt, t ≥ 0) is a Q∗ -Brownian
Girsanov’s theorem implies that the process (W
motion. Therefore,

e−rT EQ (ST 1{ST ≥K} ) = S0 Q∗ ((WT − σT ) ≥ k − σT ) (10)

³ ´
= S0 Q∗ W cT ≤ −k + σT , (11)

i.e., ³ ³ x ´´
e−rT EQ (ST 1{ST ≥K} ) = S0 N d1 .
Ke−rT
The Greeks:
The greeks for a European call option are given by:

∂CE
∆ = = N(d1) > 0
∂x
∂ 2 CE ϕ(d1)
Γ = = √ >0
∂x2 xσ T − t
∂CE
ρ = = K(T − t)e−r(T −t)N(d2) > 0
∂r
∂CE xϕ(d1)σ
Θ = =− √ − rKe−r(T −t)N(d2) < 0
∂t 2 T −t
∂CE √
V = = xϕ(d1) T − t > 0
∂σ
x2
where ϕ(x) = √12π e− 2 is the density for a standard normal
random variable.
4