• Home
  • Textbooks
  • Stochastic Calculus for Finance II : Continuous-Time Models
  • Exotic Options

Stochastic Calculus for Finance II : Continuous-Time Models

Steven E. Shreve

Chapter 7

Exotic Options - all with Video Answers

Educators


Chapter Questions

Problem 1

Exercise 7.1 (Black-Scholes-Mertonequation for the up-and-out call). This exercise shows by direct calculation that the function $v(t, x)$ of $(7.3 .20)$ satisfies the Black-Scholes-Merton equation (7.3.4).
(i) Recall that $\tau=T-t$, so $\frac{d r}{d t}=-1$. Show that $\delta_{ \pm}(\tau, s)$ given by (7.3.18) satisfies
$$
\frac{\partial}{\partial t} \delta_{ \pm}(\tau, s)=-\frac{1}{2 \tau} \delta_{ \pm}\left(\tau, \frac{1}{s}\right) \text {. }
$$
(ii) Show that for any positive constant $c$,
$$
\frac{\partial}{\partial x} \delta_{ \pm}\left(\tau, \frac{x}{c}\right)=\frac{1}{x \sigma \sqrt{\tau}}, \quad \frac{\partial}{\partial x} \delta_{ \pm}\left(\tau, \frac{c}{x}\right)=-\frac{1}{x \sigma \sqrt{\tau}} .
$$
(iii) Show that
$$
\frac{N^{\prime}\left(\delta_{+}(\tau, s)\right)}{N^{\prime}\left(\delta_{-}(\tau, s)\right)}=\frac{e^{-r \tau}}{s}
$$
and hence
$$
e^{-r \tau} N^{\prime}\left(\delta_{-}(\tau, s)\right)=s N^{\prime}\left(\delta_{+}(\tau, s)\right)
$$
(iv) Show that
$$
\frac{N^{\prime}\left(\delta_{ \pm}(\tau, s)\right)}{N^{\prime}\left(\delta_{ \pm}\left(\tau, s^{-1}\right)\right)}=s^{-\left(\frac{2 r}{\sigma^2} \pm 1\right)}
$$
and hence
$$
N^{\prime}\left(\delta_{ \pm}\left(\tau, s^{-1}\right)\right)=s^{\frac{2 r}{\sigma^2} \pm 1} N^{\prime}\left(\delta_{ \pm}(\tau, s)\right) .
$$
(v) Show that
$$
\delta_{+}(\tau, s)-\delta_{-}(\tau, s)=\sigma \sqrt{\tau} .
$$
(vi) Show that
$$
\delta_{ \pm}(\tau, s)-\delta_{ \pm}\left(\tau, s^{-1}\right)=\frac{2}{\sigma \sqrt{\tau}} \log s
$$
(vii) Show that
$$
N^{\prime \prime}(y)=-y N^{\prime}(y) .
$$

Check back soon!
01:27

Problem 2

Exercise 7.2 (Boundary conditions for the up-and-out call). In this exercise, it is verified that the up-and-out call price $v(t, x)$ given by $(7.3 .20)$ satisfies the boundary condition (7.3.6). Furthermore, the limit as $x \downarrow 0$ satisfies (7.3.5) and the limit as $t \uparrow T$ satisfies (7.3.7).
(i) Verify by direct substitution into (7.3.20) that (7.3.6) is satisfied.
(ii) Show that, for any positive constant $c$,
$$
\lim _{x \downarrow 0} \delta_{ \pm}\left(\tau, \frac{x}{c}\right)=-\infty, \lim _{x \downarrow 0} \delta_{ \pm}\left(\tau, \frac{c}{x}\right)=\infty .
$$

Use this to show that for any $p \in \mathbb{R}$ and positive constants $c_1$ and $c_2$, we have
$$
\begin{aligned}
\lim _{x \downarrow 0} x^p\left[N\left(\delta_{ \pm}\left(\tau, \frac{x}{c_1}\right)\right)-N\left(\delta_{ \pm}\left(\tau, \frac{x}{c_2}\right)\right)\right] & =0, \\
\lim _{x \downarrow 0} x^p\left[N\left(\delta_{ \pm}\left(\tau, \frac{c_1}{x}\right)\right)-N\left(\delta_{ \pm}\left(\tau, \frac{c_2}{x}\right)\right)\right] & =0 .
\end{aligned}
$$

If $p \geq 0,(7.8 .12)$ and (7.8.13) are immediate consequences of (7.8.11). However, if $p<0$, one should first use L'Hôpital's rule and then show that
$$
\lim _{x \downarrow 0} x^p \exp \left\{-\frac{1}{2} \delta_{ \pm}^2\left(\tau, \frac{x}{c_i}\right)\right\}=0, \lim _{x \downarrow 0} x^p \exp \left\{-\frac{1}{2} \delta_{ \pm}^2\left(\tau, \frac{c_i}{x}\right)\right\}=0 .
$$

To establish (7.8.14), you may wish to prove and use the inequality
$$
\frac{1}{2} a^2-b^2 \leq(a+b)^2 \text { for all } a, b \in \mathbb{R} .
$$

Conclude that $\lim _{x \downarrow 0} v(t, x)=0$ for $0 \leq t<T$.
(iii) Show that, for any positive $c$,
$$
\lim _{\tau \downarrow 0} \delta_{ \pm}(\tau, c)= \begin{cases}-\infty & \text { if } 0<c<1 \\ 0 & \text { if } c=1 \\ \infty & \text { if } c>1\end{cases}
$$

Use this to show that $\lim _{\tau \downarrow 0} v(t, x)=(x-K)^{+}$for $0<x<B$.

Ahmed Ibrahim
Ahmed Ibrahim
Numerade Educator

Problem 3

(Markov property for geometric Brownian motion and its maximum to date). Recall the geometric Brownian motion $S(t)$ of (7.4.1) and its maximum-to-date process $Y(t)$ of (7.4.3). According to Definition 2.3.6, in order to show that the pair of processes $(S(t), Y(t))$ is Markov, we must show that whenever $0 \leq t \leq T$ and $f(x, y)$ is a function, there exists another function $g(x, y)$ such that
$$
\mathbb{E}[f(S(T), Y(T)) \mid \mathcal{F}(t)]=g(S(t), Y(t)) .
$$

Use the Independence Lemma, Lemma 2.3.4, to show that such a function $g(x, y)$ exists.
7.8 Exercises
335

Check back soon!

Problem 4

4 (Cross variation of geometric Brownian motion and its maximum to date). Let $S(t)$ be the geometric Brownian motion (7.4.1) and let $Y(t)$ be the maximum-to-date process (7.4.3). Let $T$ be fixed and let $0=t_0<t_1<\ldots t_m=T$ be a partition of $[0, T]$. Show that as the number of partition points $m$ approaches infinity and the length of the longest subinterval $\max _{j=1, \ldots, m} t_j-t_{j-1}$ approaches zero, the sum
$$
\sum_{j=1}^m\left(Y\left(t_j\right)-Y\left(t_{j-1}\right)\right)\left(S\left(t_j\right)-S\left(t_{j-1}\right)\right)
$$
has limit zero.

Check back soon!

Problem 5

(Black-Scholes-Merton equation for lookback option). We wish to verify by direct computation that the function $v(t, x, y)$ of $(7.4 .35)$ satisfies the Black-Scholes-Merton equation (7.4.6). As we saw in Subsection 7.4.3, this is equivalent to showing that the function $u$ defined by (7.4.36) satisfies the Black-Scholes-Merton equation (7.4.18). We verify that $u(t, z)$ satisfies (7.4.18) in the following steps. Let $0 \leq t<T$ be given, and define $\tau=T-t$.
(i) Use (7.8.1) to compute $u_t(t, z)$, and use (7.8.3) and (7.8.4) to simplify the result, thereby showing that
$$
\begin{gathered}
u_t(t, z)=r e^{-r \tau} N\left(-\delta_{-}(\tau, z)\right)-\frac{1}{2} \sigma^2 e^{-r \tau} z^{1-\frac{2 \tau}{\sigma^2}} N\left(-\delta_{-}\left(\tau, z^{-1}\right)\right) \\
-\frac{\sigma z}{\sqrt{\tau}} N^{\prime}\left(\delta_{+}(\tau, z)\right) .
\end{gathered}
$$
(ii) Use (7.8.2) to compute $u_z(t, z)$, and use (7.8.3) and (7.8.4) to simplify the result, thereby showing that
$$
\begin{aligned}
u_z(t, z)=(1+ & \left.\frac{\sigma^2}{2 r}\right) N\left(\delta_{+}(\tau, z)\right) \\
& +\left(1-\frac{\sigma^2}{2 r}\right) e^{-r \tau} z^{-\frac{2 r}{\sigma^2}} N\left(-\delta_{-}\left(\tau, z^{-1}\right)\right)-1 .
\end{aligned}
$$
(iii) Use (7.8.19) and (7.8.2) to compute $u_z(t, z)$, and use (7.8.3) and (7.8.4) to simplify the result, thereby showing that
$$
u_{z z}(t, z)=\left(1-\frac{2 r}{\sigma^2}\right) e^{-r \tau} z^{-\frac{2 r}{\sigma^2-1}} N\left(-\delta_{-}\left(\tau, z^{-1}\right)\right)+\frac{2}{z \sigma \sqrt{\tau}} N^{\prime}\left(\delta_{+}(\tau, z)\right) .
$$
(iv) Verify that $u(t, z)$ satisfies the Black-Scholes-Merton equation (7.4.18).
(v) Verify that $u(t, z)$ satisfies the boundary condition (7.4.20).

Check back soon!

Problem 6

(Boundary conditions for lookback option). The lookback option price $v(t, x, y)$ of $(7.4 .35)$ must satisfy the boundary conditions
336
7 Exotic Options
(7.4.7)-(7.4.9). As we saw in Subsection 7.4.3, this is equivalent to the function $u(t, z)$ of $(7.4 .16)$ given by $(7.4 .36)$,
$$
\begin{aligned}
u(t, z)=\left(1+\frac{\sigma^2}{2 r}\right) z N\left(\delta_{+}(\tau, z)\right)+e^{-r \tau} N\left(-\delta_{-}(\tau, z)\right) & \\
& -\frac{\sigma^2}{2 r} e^{-r \tau} z^{1-\frac{2 r}{\sigma^2}} N\left(-\delta_{-}\left(\tau, z^{-1}\right)\right)-z, 0 \leq t<T, 0<z \leq 1,
\end{aligned}
$$
satisfying the boundary conditions (7.4.19)-(7.4.21). This function was shown to satisfy boundary condition (7.4.20) in Exercise 7.5(v). Here we verify by direct computation that the limit of $u(t, z)$ as $z \downarrow 0$ satisfies (7.4.19) and the limit of $u(t, z)$ as $t \uparrow T(\tau \downarrow 0)$ satisfies (7.4.21).

Check back soon!

Problem 7

(Zero-strike Asian call). Consider a zero-strike Asian call whose payoff at time $T$ is
$$
V(T)=\frac{1}{T} \int_0^T S(u) d u .
$$
(i) Suppose at time $t$ we have $S(t)=x \geq 0$ and $\int_0^t S(u) d u=y \geq 0$. Use the fact that $e^{-r u} S(u)$ is a martingale under $\widetilde{\mathbb{P}}$ to compute
$$
e^{-r(T-t)} \tilde{\mathbb{E}}\left[\frac{1}{T} \int_0^T S(u) d u \mid \mathcal{F}(t)\right] .
$$

Call your answer $v(t, x, y)$.
(ii) Verify that the function $v(t, x, y)$ you obtained in (i) satisfies the BlackScholes-Merton equation (7.5.8) and the boundary conditions (7.5.9) and (7.5.11) of Theorem 7.5.1. (We do not try to verify $(7.5 .10)$ because the computation of $v(t, x, y)$ outlined here works only for $y \geq 0$.)
(iii) Determine explicitly the process $\Delta(t)=v_x(t, S(t), Y(t))$, and observe that it is not random.
(iv) Use the Itô-Doeblin formula to show that if you begin with initial capital $X(0)=v(0, S(0), 0)$ and at each time you hold $\Delta(t)$ shares of the underlying asset, investing or borrowing at the interest rate $r$ in order to do this, then at time $T$ the value of your portfolio will be
$$
X(T)=\frac{1}{T} \int_0^T S(u) d u
$$

Check back soon!
01:16

Problem 8

Consider the continuously sampled Asian option of Subsection 7.5.3, but assume now that the interest rate is $r=0$. Find an initial capital $X(0)$ and a nonrandom function $\gamma(t)$ to replace (7.5.22) so that
$$
X(T)=\frac{1}{c} \int_{T-c}^T S(u) d u-K
$$
still holds. Give the formula for the resulting process $X(t), 0 \leq t \leq T$, to replace (7.5.24) and (7.5.26). With this function $\gamma(t)$ and process $X(t)$, Theorem 7.5.3 still holds.

Benjamin Chaback
Benjamin Chaback
Numerade Educator
01:03

Problem 9

. Let $g(t, y)$ be the function in Theorem 7.5.3. Then the value of the Asian option at time $t$ is $V(t)=v(t, S(t), X(t))$, where $v(t, s, x)=s g(t, y)$ and $y=\frac{x}{s}$. The process $S(t)$ is given by (7.5.1). For the sake of specificity, we consider the case of continuous sampling with $r \neq 0$, so $\gamma(t)$ is given by (7.5.22) and $X(t)$ is given by (7.5.24) and (7.5.26).
(i) Verify the derivative formulas

Hoan Nguyen
Hoan Nguyen
Numerade Educator