Summer 2022 Exam

MA322
Mathematics of Finance and Valuation

Suitable for all candidates

Instructions to candidates

This paper contains 4 questions. Your answers to all 4 questions will count towards the final mark.
All questions carry an equal number of marks.
Answers should be justified by showing work.
Please write your answers in dark ink (black or blue) only.

Time Allowed Reading Time: None

Writing Time: 2 hours

You are supplied with: No additional materials

You may also use: No additional materials

Calculators: Calculators are not allowed in this exam.

© LSE ST 2022/MA322 Page 1 of 5

Question 1
In this question W denotes a Brownian motion, while µ, σ , and T are strictly positive constants. You
may assume, without proof, that all integrals with respect to the Brownian Motion are martingales.
(a) Consider the stochastic differential equation
µ
dLt = − Lt dWt , L0 = 1, (1)
σ(1 + t)
on the interval [0, T ].
(i) Verify that the solution of equation (1) is given by
t
µ2 t
 Z 
µ 1
Lt = exp − dWs − 2 .
σ 0 1+s 2σ (1 + t)

(ii) Calculate the mean and the variance of log(Lt ).

(b) (i) Derive the solution of the stochastic differential equation

1
dYt = (µ − Yt ) dt + σdWt , Y0 = 0. (2)
1+t

(ii) Prove that the process defined by

Mt = Lt (1 + t)Yt ,

where the dynamics of L are given by (1) and those of Y by (2), is a martingale.

© LSE ST 2022/MA322 Page 2 of 5

Question 2
(a) Give the definition of the Brownian motion on [0, T ].
(b) Let W 1 and W 2 be two independent standard Brownian motions on [0, T ]. Consider the process

Wt1 + Wt2
Xt = √ , t ∈ [0, T ].
2
Is X a Brownian motion? Justify your answer by either proof or counterexample, as appropriate.
Hint. You can use without proof the following facts: if {Xi }4i=1 are independent normal random
P4
variables and {ai }ni=1 , {bi }n
i=1 are real numbers, then the random variables i=1 ai Xi and
P4
i=1 bi Xi are jointly normal; if X and Y are jointly normal, then they are independent if and
only if they are uncorrelated.
(c) Let W be a standard Brownian motion and consider the stochastic differential equation

Xt
dXt = − dt + dWt , Y0 = 0 (3)
1−t
on the interval [0, 1).
Is the stochastic process Yt = (1 + t)X t , t ≥ 0 a Brownian motion on an interval [0, T ] for
1+t
any T > 0? Justify your answer by either proof or counterexample, as appropriate.
Hint. You may use the fact that for any deterministic square integrable function f on [0, T ] and
Rt Rt
0 ≤ t1 < t2 < · · · < tk ≤ T the random variables t 2 f (u)dWu , . . . , t k f (u)dWu are
1 k−1
independent.
(d) Consider an adapted integrable stochastic process M on a filtered probability space (Ω, F, (Ft ), P)
satisfying E[Mt ] = E[Ms ] for all s, t ∈ [0, T ]. Is the process M a martingale on [0, T ]? Justify
your answer by either proof or counterexample, as appropriate.

© LSE ST 2022/MA322 Page 3 of 5

Question 3
Throughout this question, assume a standard Black and Scholes market consisting of a risk-free
money market account satisfying

dBt = rBt dt,

and a stock with price process S given by

dSt = µSt dt + σSt dWt ,

where µ, σ , and r ≥ 0 are constants. Consider a portfolio from which funds are withdrawn at a rate ct
at time t. (The process c is assumed to be bounded and adapted to the market filtration.) Thus, the
Rt
total amount of payments made between time s ≥ 0 and t > s is s cu du. The wealth process X of
such a portfolio is given by

Xt = Nt0 Bt + Nt St which has the dynamics dXt = Nt0 dBt + Nt dSt − ct dt,

where Nt0 denotes the units of the money market account and Nt the number of shares held at time
t.
(a) Fix T > 0. Show that
Z T Z T
XT −ru
= X0 + σe πu dWuQ − cu e−ru du,
BT 0 0

where W Q is a Brownian motion under the risk-neutral measure Q and πt is the amount of
money invested in the stock at time t.
(b) Let T , π ∈ H2 and X be as in (a). Show that if XT = 0, then X0 is the no-arbitrage price of the
derivative which makes continuous payments at a rate c over the time period 0 to T .
(c) Consider a derivative that yields to its holder a stream of payment at rate 1 up to the random
time τa given by

τa = inf t ≥ 0 | St = a , a ≥ S0 .

The total amount of payments that such a derivative makes to its holder between time s ≥ 0
and time t > s is t ∧ τa − s ∧ τa , where a ∧ b = min{a, b}. (With the notation in (a) this
corresponds to c = 1 and T = τa .) Denote by Pt the no-arbitrage price of this derivative at time
t ∈ [0, +∞).
(i) Derive a partial differential equation with suitable boundary conditions that a C 2 function
s 7→ v(s) should satisfy so that

Pt = v(St )1{t<τa } . (4)

Hint. You may use the fact that that τa is a stopping time that is (almost surely) finite. You
may also assume that all stochastic integrals with respect to W Q are martingales under
the risk-neutral measure Q.
(ii) State an expression for the hedging portfolio process that is associated with equality (4).

© LSE ST 2022/MA322 Page 4 of 5

Question 4
(a) Throughout this part of the question, assume a standard Black and Scholes market consisting
of a risk-free money market account satisfying

dBt = rBt dt,

and a stock with price process S given by
  
1 2 Q
St = S0 exp r − σ t + σWt ,
2
where r ≥ 0 and σ are constants and W Q is a standard Brownian motion under the risk neutral
F (S )
probability measure Q. Also, recall that the no-arbitrage price Pt T of a European contingent
claim that yields a payoff F (ST ) at its maturity time T is given by
F (ST )
Pt = e−rT EQ [F (ST ) | Ft ] , for t ∈ [0, T ].

For p, K > 0, let C0 (T, K, p) denote the price at time 0 of the European power call option with
payoff

+
F (ST ) = STp − K
at time T .
p
(i) Set Yt = St for t ≥ 0. Show that the process (Yt ) satisfies the SDE

dYt = γYt dt + pσYt dWtQ , Y0 = S0p

for some constant γ that you should specify.
(ii) Derive an analytic expression for the price C0 (T, K, p).
Hint. You may use without proof that for a random variable Z ∼ N (0, 1) under Q we have
λ2
h i
EQ 1{Z≥a} = Q (Z ≥ a) = Φ(−a) and EQ eκ+λZ 1{Z≥a} = eκ+ 2 Φ(λ − a),
 

where a, κ, λ are constants and Φ is the standard normal distribution function, given by
c
u2
Z
1
Φ(c) = √ e− 2 du, for c ∈ R.
2π −∞

(b) Consider a market with finite horizon T > 0 in which there are liquidly traded calls for all strikes
and maturities with market prices (at time zero) given by C M KT (u, K), u ∈ [0, T ]. Suppose
that the interest rate is zero and calibrated local volatility σ(t, s) = σ(t), where σ : t 7→ σ(t) is
a continuously differentiable strictly positive function.
In this setting, the market call prices match the call prices calculated under the assumption that
the stock price process S satisfies

dSt = σ(t)St dWtQ ,

where W Q is a standard Brownian motion under the risk neutral probability measure Q (you are
not required to prove this).
Is it true that if the implied volatility at time zero, Σ(t, K), satisfies Σ(t, K) = σ(t) for all K > 0
and all t ∈ [0, T ], then Σ(t, K) = σ(t) = σ for some constant σ ∈ [0, +∞)? Justify your
answer by proof or counterexample, as appropriate.

END OF PAPER

© LSE ST 2022/MA322 Page 5 of 5