MAT00028M Stochastic Calculus and Black-Scholes Theory 2018/19
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
MAT00028M
MMath and MSc Examinations 2018/19
Stochastic Calculus and Black-Scholes Theory
1 (of 3). Let (Ω , F, P) be a probability space, let (Ft )to0 be a filtration on this space
and let W be a Brownian Motion with respect to (Ft )to0 .
(a) Define what is meant by Brownian motion. Also, define what is meant
(c) Show whether ξ , η defined by
5, if t e [0, 6) 4, if t e [0, 3)
ξ(t) = , η(t) = 2W(W2) ,
(0, if t > 10 (0, if t > 9
are random step processes. Give reasons for your answer. If any of these processes is indeed a random step process, then calculate the mean and
variance of its stochastic integral I(.). [7]
(d) (i) State the Itˆo Lemma in either its simple or general form.
(ii) Use the Itˆo lemma to show that the stochastic process Z defined by Zt = Wt2 _ t, t > 0 is a martingale.
(iii) Use the previous part of the question to determine the quadratic variation process (W) of the Brownian motion W . Show that the process found by you satisfies the definition of the quadratic varia- tion of a continuous square-integrable martingale. [8]
Let (Ω , F, P) be a probability space, let (Ft )to0 be a filtration on this space and let W be a Brownian Motion with respect to (Ft )to0 .
(a) (i) Define the spaces M2 (0, o) and Mlo(2)c (0, o).
(ii) Show that the process A defined by At = sin (exp (7Wt3 )) belongs to Mlo(2)c (0, o).
(iii) Show that the Brownian motion W is not in M2 (0, o). [6]
(b) Show that the stochastic process X defined by Xt = sin (Wt ) satisfies the stochastic differential equation
dXt = cos (Wt ) dWt _ Xt dt, X0 = 0.
[4]
(c) Let X and Y be Itˆo processes with characteristics aX , bX and aY , bY respectively, i.e. suppose that
dXt = aX (t) dt + bX (t) dWt
and
dYt = aY (t) dt + bY (t) dWt .
Prove the stochastic integration by parts formula
d (XY)t = Xt dYt + Yt dXt + bX (t) bY (t) dt.
You may assume without proof that the processes X _ Y and X + Y are
Itˆo processes. [6]
(d) Consider the stochastic differential equation
dXt = (Xt )γ dt + βXt dWt , X0 = x > 0,
where 0 < γ s 1, β > 0 are constants. Consider the process G defined by
Gt = exp ╱ _βWt + β 2t、
and let Zt = GtXt , t > 0. Show that Z satisfies
dZt = (Zt )γ (Gt )1_γ dt, Z0 = x, (1)
which is an ordinary differential equation (ODE). For 0 < γ < 1 solve this ODE for Zt and, hence, find the solution Xt . State the solution Xt if γ = 1 (you do not need to prove your result for the particular case γ = 1).
Hint: Apply the Itˆo formula to the Itˆo process G to determine its char- acteristics. Next, use stochastic integration by parts for the product
GtXt in order to derive equation (1). [9]
3 (of 3). Suppose that σ > 0, r > 0, T > 0 and µ e R are fixed. Let (Ω , F, P) be
a probability space and let W be a Brownian motion defined on this space. Consider the Black-Scholes model, i.e. a market consisting of a stock and a bond with prices given by stochastic processes S and B respectively. Assume that S and B are solutions to the following stochastic differential equations:
dSt = µSt dt + σSt dWt , t > 0,
dBt = rBt dt, t > 0.
(a) Let (ϕ, ψ) be a trading strategy in the Black-Scholes model.
(i) Define the wealth process associated to the trading strategy (ϕ, ψ). Furthermore, define what is meant by a self-financing trading strat- egy.
(ii) Show that a trading strategy (ϕ, ψ) is self-financing if and only if its wealth process X satisfies
dXt = (rXt + (µ _ r) ϕt St ) dt + σϕt St dWt , t > 0. [5]
(b) State the Black-Scholes partial differential equation (PDE). [2]
(c) Consider a European option whose payoff is h (ST ) at time T. In the following assume that its value C (t, x) at time t < T is a C1,2 function of t and x. Suppose that the processes ϕ, ψ are defined by
∂C (t, St ) + σ 2 St(2) (t, St )
Suppose also that X is the wealth process corresponding to the trading strategy (ϕ, ψ) and C satisfies the Black-Scholes PDE with terminal condition
C (T, x) = h (x) .
Show that the wealth process X satisfies
C (t, St ) = Xt , t e [0, T]
and (ϕ, ψ) is a self-financing strategy.
[4]
(d) Define the risk neutral measure Q in the context of the Black-Scholes model. Give an explicit expression for , i.e. the Radon-Nikodym derivative of the risk neutral measure with respect to the physical mea- sure (no proof required). [3]
(e) State and prove the put-call parity theorem. [5]
(f) (i) State the theorem on the replicability of options in the Black-Scholes model (no proof required).
(ii) An Asian put option is a put option with exercise time T > 0 and strike K > 0 on the average price of the stock from time 0 to time T. What is the payoff for this option? Prove that the Asian put option is replicable in the Black-Scholes model. You may use the theorem on the replicability of options provided you show its assumptions are satisfied.
[6]
(a) A Brownian Motion is a stochastic process (Wt )to0 on a probability space (Ω , F, P) such that
1. W0 = 0 a.s.
2. for all k e N and all 0 = t0 < t1 < . . . < tk the r.v.s Wt1 _ Wt0 , Wt2 _ Wt1 , . . . , Wtk _ Wtk − 1 are independent (independent increments)
3.Wt _ Ws ~ N(0, t _ s) for all t > s > 0
4.Trajectories are continuous a.s.
We say that (Wt )to0 is a Brownian Motion with respect to the filtration (Ft )to0 (or (Ft )to0-Brownian motion) if:
1. the random variable Wt is Ft-measurable for every t > 0,
2. the random variable Wt _ Ws is independent of Fs for all 0c、.
(b) If ξ e Mst(2)ep (0, o), then
we can find N e N and a sequence of times
0 = t0 < t1 < . . . < tN
such that
ξ(t) =
Hence, we have on the one hand
E|I(ξ)| 2 = E ┌ ξ(ti )(W (ti+1) _ W (ti ))!2
= E ┌ξ(ti )2 (W (ti+1) _ W (ti ))2 ┐
+ 2 z E ┌ξ(ti )ξ(tj )[W (ti+1) _ W (ti )][W (tj+1) _ W (tj 2]┐)
Since ξ(ti )2 is Fti _measurable and W (ti+1) _ W (ti ) is independent of Fti we have for i = 0, . . . , N _ 1
┌ ┐ ┌ ┐ ┌ ┐
= E ┌ξ(ti )2 ┐ (ti+1 _ ti ).
For the terms in the second sum we note that the random variables
ξ(ti ), W (ti+1) and W (ti ) are Ftj _measurable and W (tj+1) _ W (tj ) is
SOLUTIONS: MAT00028M
independent of Ftj . Hence,
E ┌ξ(ti )ξ(tj )[W (ti+1) _ W (ti )][W (tj+1) _ W (tj )]┐
┌ ┐
Combining the two calculations we see that
E|I(ξ)| 2 = E ┌ξ(ti )2 ┐ (ti+1 _ ti ).
On the other hand, note that
E |ξt | 2 =
and therefore
ì0 o E|ξt | 2 dt
= E|ξt0 | 2 (t1 _ t0 ) + E|ξt1 | 2 (t2 _ t1 ) + . . . + E|ξtN − 1 | 2 (tN _ tN _1 ).
c 、
7 Marks
s '
(c) ξ is not a random step process as it is not adapted (6W13 is not F8 measurable). For η we check that W2 is F3 measurable and 2W4(2) is F6 measurable, the constants are measurable anyway, hence η is adapted. Noting that E (W4(4)) < o and E(W2 )2 < o (moments of normal random variables) the integrability condition also holds. The step property holds for the partition 0 < 3 < 6 < 9 and the process vanishes for t > 9. Hence, η is a random step process. Finally, E (I (η)) = 0 either by general prop-
erties of the stochastic integral or by direct calculation. Hence, using the
Ito Isometry Var(I (η)) = E ╱I (η)2 、= 16 × 3 + 2 × 3 + 4EW4(4) × 3 =
s '
(d) (i) Theorem 0.1 |上tao 上emmb in simple form女
与ssume thbt f (t, x) is in C1 2, (R )! i冬e冬 the derivbtives2 , ,
ezist bnd bre continuous冬 Ωe忐ne b process η 6y
η(t) = f (t, W (t)), t > 0.
7hen η(t)! t > 0 is bn 上tao process bnd
η(t) = f (0, 0) + ì0 t [ (s, W (s)) + (s, W (s))]ds + ì0 t (s, W (s)) dW (s), t > 0.
(ii) We apply the the Ito lemma in its simple form with f (t, x) = x2 . Note that f e C1,2 and fx (t, x) = 2x, fxx (t, x) = 2, ft = 0. Hence,
Wt2 = ì0 t 2Ws dWs + ì0 t dt
and Wt2 _ t can be written as a stochastic integral. It is easy to check that the integrand 2Ws is in Mlo(2)c (0, o) . Adaptedness and continuity is satisfied by definition of the BM. Moreover, E(4Ws )2 = 4s so the integrability condition holds as well. We deduce that the stochastic integral is a martingale.
(iii) For the final part we note that the process (W) given by (W)t = t, t > 0 is an adapted, continuous and increasing process with (W) 0 = 0. Moreover, Wt2 _ (W)t is a martingale by the previous part of the question. Hence, (W) satisfies the defintion of the quadratic
variation of the square-integrable martingale W.
c 、
8 Marks
s '
c 、
Total: 25 Marks
2. (a) We say that a process (ξt )to0 belongs to class M2 (0, o) if and only if
(i) (ξt )to0 is adapted to the filtration (Ft )to0 .
(ii) the trajectories of (ξt )to0 are left continuous or right continuous a.s. (iii) E |0o |ξt | 2 dt < o.
We say that a process (ξt )to0 belongs to class Mlo(2)c (0, o) if and only if (i) (ξt )to0 is adapted to the filtration (Ft )to0 .
(ii) the trajectories of (ξt )to0 are left continuous or right continuous a.s. (iii) For every T > 0, E |0T |ξt | 2 dt < o.
Note that the process A is a continuous function of Wt and Wt is BM (which is adapted and has by definition a.s. continuous sample paths). Hence, we can deduce that At is adapted with continuous sample paths. It remains to check the integrability condition. For every T > 0 using that the cosine function is bounded above by one we have
ì0 T E ╱sin2 ╱exp(7(Wt )3 )、、dt s ì0 T dt = T < o.
Brownian motion is not in Mlo(2)c (0.o) as E |0o |Wt | 2 dt = |0o c(tdt) = o.、
6 Marks
s '
(b) If f (x) = sin (x) we have = cos (x) , = _ sin (x) and = 0, so f e C1,2 . Hence, we can apply Ito’s formula and get
dXt = _ sin (Wt ) dt + cos (Wt ) dWt .
Substituting Xt for sin (Wt ) in the above equation and observing that
sin (W0 ) = 0 we see that Xt solves the SDE as required.
c 、
4 Marks
s '
(c) Applying the Ito lemma in its general form with f (t, x) = x2 to the Ito processes (X + Y) and (X _ Y) yields
d ╱(X + Y)2、t = ╱(aX (t) + aY (t)) 2 (Xt + Yt )) + (bX (t) + bY (t))2、dt + (bX (t) + bY (t)) 2 (Xt + Yt ) dWt
and
d ╱(X _ Y)2、t = ╱(aX (t) _ aY (t)) 2 (Xt _ Yt )) + (bX (t) _ bY (t))2、dt + (bX (t) _ bY (t)) 2 (Xt _ Yt ) dWt
Subtracting the second from the first equation and dividing by four we see that
d (XY)t = [4aX (t) Yt + 4aY (t) Xt + 4bX (t) bY (t)] dt + (4bX (t) Yt + 4bY (t) Xt ) dWr
= aX (t) Yt dt + bX (t) Yt dWt + aY (t) Xt dt + bY (t) Xt dWt
+bX (t) bY (t) dt
= Yt dXt + Xt dYt + bX (t) bY (t) dt.
c 、
6 Marks
s '
(d) Note that the function
f (t, x) = exp ╱ _βx + β 2t、
is in C1,2 . Hence, we can apply the Itˆo lemma to Gt = f (t, Wt ) and deduce that
dGt = β 2 Gt dt + β 2 Gt dt _ βGt dWt
= β 2 Gt dt _ βGt dWt .
The stochastic integration by parts formula for two processes dξt = a1 (t) dt + b1 (t) dWt and dηt = a2 (t) dt + b2 (t) dWt states that
d (ξt ηt ) = ξt dηt + ηt dξt + b1 (t) b2 (t) dt.
Applying stochastic integration by parts to Zt = Xt Gt we have
dZt = Xt dGt + Gt dXt _ β 2 GtXt dt
= Xt β 2 Gt dt _ Xt βGt dWt + Gt (Xt )γ dt + Gt βXt dWt _ β 2 GtXt dt = β 2 Zt dt _ βZt dWt + Gt (Xt )γ dt + βZt dWt _ β 2 Zt dt
= Gt (Xt )γ dt
= (Gt )1_γ (Zt )γ dt.
We now solve this ODE for 0 < γ < 1 by considering
dZt = Zt(γ)Gt(1) _γ dt, Z0 = 0.
Using separation of variables
ì0 t = ì0 t Ft1_γ dt
and hence,
Z x1_γ = (1 _ γ) ì0 t Ft1_γ dt.
Therefore,
1
Zt = ╱x1_γ + (1 _ γ) ì0 t Ft1_γ dt、1 −γ .
Finally, if γ = 1 we know from the examples in the lectures that Xt is the stochastic exponential
Xt = x0 exp ╱╱1 _ β 2 /2、t + βWt、.
c 、
9 Marks
s '
c 、
Total: 25 Marks
3. (a) (i) The wealth process X(t), t > 0 corresponding to the trading strategy (ϕ, ψ), is defined by
X(t) := ψ(t)B(t) + ϕ(t)S(t), t > 0. (3)
A trading strategy (ϕ, ψ) is self-financing if the corresponding wealth process satisfies
dXt = ψ(t) dBt + ϕ(t) dSt , t > 0 (4)
or equivalently
ψt Bt + ϕt St _ (ψs Bs + ϕs Ss ) = ϕu dSu + ψu dBu
s s
for s, t > 0.
(ii) The claim follows from the equality
(rXt + (µ _ r) ϕt St ) dt + σϕt St dWt = (r (ϕt St + ψt Bt ) + (µ _ r) ϕt St ) dt + σϕt St dWt
= (rψt Bt + µϕt St ) dt + σϕt St dWt = ψtrBt dt + ϕt (µSt dt + σSt dWt ) = ψt dBt + ϕt dSt.
c 、
5 Marks
s '
(b) The Black-Scholes PDE is given by
+ rx + σ 2 x2 _ rC = 0, t e [0, T] , x > 0. (5)
c 、
2 Marks
s '
(c) The wealth process corresponding to (ϕ, ψ) satisfies
∂C (t, St ) + σ 2 St(2) (t, St )
∂x rBt
where we have used that C satisfies the Black-Scholes PDE in the last step. To show that the self-financing property holds we note that using the Ito formula
ì0 t ϕs dSs + ì0 t ψs dBs = ì0 t (t, Ss ) dSt
+ ì0 t ╱ (s, Ss ) + σ S2s(2) (t, Ss )、ds
= C (t, St ) _ C (0, S0 ) .
c 、
4 Marks
s '
(d) The risk neutral measure Q in the B-S model is the unique P equivalent measure under which the stock price discounted by the risk free interest
rate is a martingale. In this case we have (restricting the measures to FT )
= exp ╱ _ WT _ T\
c 、
3 Marks
s '
(e) Put-Call parity theorem:
Let T > 0 and let C(t), t e [0, T], be the value of a call option with strike price K and maturity time T. Let P (t) be the value of a put option with the same strike and maturity. We denote by S(t) the price of the share at time t, and r the risk-free interest rate. Then
C(t) _ P (t) = S(t) _ Ker(t_T) .
(6)
Proof: Suppose that at time t = 0 we buy a call option and sell a put option with the same maturity time T and strike price K . At time T, the payoff for this portfolio is S(T) _ K, as
C(T) _ P (T) = |S(T) _ K|+ _ |K _ S(T)|+ = S(T) _ K.
Also at time 0, we can can buy a unit of share and borrow M :=
Ke_rT using bonds (hence, at time T we have to return exactly erT M =
erT Ke_rT = K). The payoff for this second portfolio at time T is also
S(T) _ K . As we assume the no br6itrbge principle we conclude that
the prices of these two portfolios have to be the same for every t e [0, T],
s '
(f) (i) Theorem 0.2 Under the bssumptions of the 去lbck」scholes model
suppose thbt h is bn option with ezpiry time T > 0 冬 suppose more」 over thbt
EQT |h|2 < o,
bnd h is b non」negbtive rbndom vbrib6le b冬s冬冬 7hen the option h is replicb6le冬 koreover! if
Xt = ϕt St + ψt Bt , t e [0, T]
is the weblth process corresponding to the bdmissi6le replicbtion strbt」 egy (ϕt , ψt ) |in pbrticulbr sbtisfying XT = h)! then
Xt = EQT ╱e_r(T _t)h F│t、, t e [0, T]. (7)
2022-08-08