National University of Singapore QF3103: Advanced Mathematics in Quantitative Finance
Tutorial 3
1. On 01_general_probability_theory slides 28–29, we began with a standard normal random vari-
able X under a measure P, which by the exercise on 03_brownian_motion slide 24 has the
moment-generating function
1 2
EeuX = e 2 u for all u ∈ R.
The moment-generating function of a random variable determines its distribution. In particular,
1 2
any random variable that has moment-generating function e 2 u must be standard normal.
On 01_general_probability_theory slides 28–29, we also defined Y = X + θ , where θ is a con-
1 2
stant, we set Z = e−θ X − 2 θ , and we defined P
e by the formula
Z
e(A) =
P Z(ω) dP(ω) for all A ∈ F .
A
Prove that Y is standard normal under P
e by verifying the moment-generating function formula
1 2
e euY = e 2 u
E for all u ∈ R.
2. 03_brownian_motion slide 30 provides the Laplace transform of the density of the first passage
time for Brownian motion. This problem derives the analogous formula for Brownian motions
with drift. Let W be a standard Brownian motion. Fix m > 0 and µ ∈ R. For 0 ≤ t < ∞, define
X (t) = µt + W (t),
τm = min{t ≥ 0; X (t) = m}.
As usual, we set τm = ∞ if X (t) never reaches the level m. Let σ be a positive number and set
1 2
Z(t) = exp − σµ + σ t + σX (t) .
2
(In the following parts, you may take it for granted that the relevant convergence theorems can
be applied similarly here.)
(i) Show that Z(t), t ≥ 0, is a martingale.
(ii) Use (i) to conclude that
1 2
E exp − σµ + σ (t ∧ τm ) + σX (t ∧ τm ) = 1, t ≥ 0.
2
(iii) Now suppose µ ≥ 0. Show that, for σ > 0,
1 2
E exp − σµ + σ τm + σm 1{τm <∞} = 1.
2
Use this fact to show P{τm < ∞} = 1 and to obtain the Laplace transform
p 2
Ee−ατm = e mµ−m 2α+µ for all α > 0.
(iv) Show that if µ > 0, then Eτm < ∞. Obtain a formula for Eτm .
(Hint: Differentiate the formula in (iii) with respect to α.)
1/2
National University of Singapore QF3103: Advanced Mathematics in Quantitative Finance
(v) Now suppose µ < 0. Show that, for σ > −2µ,
1 2
E exp − σµ + σ τm + σm 1{τm <∞} = 1.
2
Use this fact to show that P{τm < ∞} = e2m|µ| , which is strictly less than 1, and to obtain
the Laplace transform
p 2
Ee−ατm = e mµ−m 2α+µ for all α > 0.
3. Let the wealth of a company at time t be modelled by X (t) = x + µt + σW (t), where x is a
positive number. In this question, we consider the case of positive drift (µ > 0). Let τ x be the
first time when the process X hits 0, τ x = min{t : X (t) = 0}. Denote the probability of ruin
Ψ(x) = P(τ x < ∞). Note that the index x in τ x refers to the starting point and not the target
value.
Let µ > 0. Show that Ψ(x) ≤ e−Rx with R = 2µ/σ2 .
(Hint: Show that M (t) = e−RX (t) is a martingale. Then stop it at t ∧ τ x and use EM (t ∧ τ x ) =
EM (0) to obtain an upper bound on P{τ x ≤ t}.)
4. (Exit of Brownian motion from an interval) Let W (t) be a standard Brownian motion, and let
τ be the first time when B(t) exists the interval (a, b), a < 0 < b, that is, τ = min{t : W (t) =
a or b}. Clearly, τ is a stopping time.
(i) Use the basic stopping equation for a stopped martingale EM (t ∧ τ) = EM (0) to obtain
EW (τ). (You may take it for granted that the dominated convergence theorem can be
applied to interchange limit and expectation for this question.)
(ii) Find the probability that standard Brownian motion reaches b before it reaches a.
(iii) By stopping the martingale W 2 (t) − t, show that Eτ = −a b.
2/2