• Home
  • Textbooks
  • Stochastic Calculus for Finance II : Continuous-Time Models
  • American Derivative Securities

Stochastic Calculus for Finance II : Continuous-Time Models

Steven E. Shreve

Chapter 8

American Derivative Securities - all with Video Answers

Educators


Chapter Questions

01:25

Problem 1

(Determination of $L$. by smooth pasting). Consider the function $v_L(x)$ in (8.3.11). The first line in formula (8.3.11) implies that the left-hand derivative of $v_L(x)$ at $x=L$ is $v_L\left(L^{-}\right)=-1$. Use the second line in formula (8.3.11) to compute the right-hand derivative $v_L^{\prime}(L+)$. Show that the smooth-pasting condition
$$
v_{L_*}^{\prime}\left(L_*-\right)=v_{L *}^{\prime}\left(L_*+\right)
$$
is satisfied only by $L *$ given by (8.3.12).

Carson Merrill
Carson Merrill
Numerade Educator

Problem 2

Consider two perpetual American puts on the geometric Brownian motion (8.3.1). Suppose the puts have different strike prices, $K_1$ and $K_2$, where $0<K_1<K_2$. Let $v_1(x)$ and $v_2(x)$ denote their respective prices, as determined in Section 8.3.2. Show that $v_2(x)$ satisfies the first two linear complementarity conditions,
$$
\begin{aligned}
v_2(x) & \geq\left(K_1-x\right)^{+} \text {for all } x \geq 0, \\
r v_2(x)-r x v_2^{\prime}(x)-\frac{1}{2} \sigma^2 x^2 v_2^{\prime \prime}(x) & \geq 0 \text { for all } x \geq 0,
\end{aligned}
$$
for the perpetual American put price with strike $K_1$ but that $v_2(x)$ does not satisfy the third linear complementarity condition:
for each $x \geq 0$, equality holds in either (8.8.1) or (8.8.2) or both.

Check back soon!

Problem 3

(Solving the linear complementarity conditions). Suppose $v(x)$ is a bounded continuous function having a continuous derivative and satisfying the linear complementarity conditions (8.3.18)-(8.3.20). This exercise shows that $v(x)$ must be the function $v_{L_{.}}(x)$ given by (8.3.13) with $L$. given by (8.3.12). We assume that $K$ is strictly positive.
(i) First consider an interval of $x$-values in which $v(x)$ satisfies (8.3.19) with equality, i.e., where
$$
r v(x)-r x v^{\prime}(x)-\frac{1}{2} \sigma^2 x^2 v^{\prime \prime}(x)=0 .
$$

Check back soon!

Problem 4

. It was asserted at the end of Subsection 8.3.3 and established in Exercise 8.3 that $v_{L .}(x)$ given by (8.3.13) is the only bounded continuous function having a continuous derivative and satisfying the linear complementarity conditions (8.3.18)-(8.3.20). There are, however, unbounded functions that satisfy these conditions. Let $0<L<K$ be given, and assume that
$$
\frac{2 r}{2 r+\sigma^2} K>L \text {. }
$$
(i) Show that, for any constants $A$ and $B$, the function
$$
f(x)=A x^{-\frac{2 r}{\sigma^2}}+B x
$$
satisfies the differential equation
$$
r f(x)-r x f^{\prime}(x)-\frac{1}{2} \sigma^2 x^2 f^{\prime \prime}(x)=0 \text { for all } x \geq 0 .
$$
372
8 American Derivative Securities
(ii) Show that the constants $A$ and $B$ can be chosen so that
$$
f(L)=K-L, \quad f^{\prime}(L)=-1 .
$$
(iii) With the constants $A$ and $B$ you chose in (ii), show that $f(x) \geq(K-x)^{+}$ for all $x \geq L$.
(iv) Define
$$
v(x)=\left\{\begin{array}{l}
K-x, 0 \leq x \leq L, \\
f(x), \quad x \geq L .
\end{array}\right.
$$

Show that $v(x)$ satisfies the linear complementarity conditions (8.3.18)$(8.3 .20)$, but $v(x)$ is not the function $v_{L .}(x)$ given by (8.3.13).
(v) Every solution of the differential equation (8.8.7) is of the form (8.8.6). In order to have a bounded solution, we must have $B=0$. Show that in order to have $B=0$, we must have $L=\frac{2 r}{2 r+\sigma^2} K$, and in this case $v(x)$ agrees with the function $v_{L_*}(x)$ of (8.3.13).

Check back soon!

Problem 5

(Perpetual American put paying dividends). Consider a perpetual American put on a geometric Brownian motion asset price paying dividends at a constant rate $a>0$. The differential of this asset is
$$
d S(t)=(r-a) S(t) d t+\sigma S(t) d \widetilde{W}(t),
$$
where $\widetilde{W}(t)$ is a Brownian motion under a risk-neutral measure $\widetilde{\mathbb{P}}$. (Equation (8.8.9) can be obtained by computing the differential in (5.5.8).)
(i) Suppose we adopt the strategy of exercising the put the first time the asset price is at or below $L$. What is the risk-neutral expected discounted payoff of this strategy? Write this as a function $v_L(x)$ of the initial asset price $x$. (Hint: Define the positive constant
$$
\gamma=\frac{1}{\sigma^2}\left(r-a-\frac{1}{2} \sigma^2\right)+\frac{1}{\sigma} \sqrt{\frac{1}{\sigma^2}\left(r-a-\frac{1}{2} \sigma^2\right)^2+2 r}
$$

Check back soon!

Problem 6

. There is a second part to Theorem 8.2.4 (optional sampling), which says the following.

Theorem 8.8.1 (Optional sampling - Part II). Let $X(t), t \geq 0$, be $a$ submartingale, and let $\tau$ be a stopping time. Then $\mathbb{E} X(t \wedge \tau) \leq \mathbb{E} X(t)$. If $X(t)$ is a supermartingale, then $\mathbb{E} X(t \wedge \tau) \geq \mathbb{E} X(t)$. If $X(t)$ is a martingale, then $\mathbb{E} X(t \wedge \tau)=\mathbf{E} X(t)$.

The proof is technical and is omitted. The idea behind the statement about submartingales is the following. Submartingales tend to go up. Since $t \wedge \tau \leq t$, we would expect this upward trend to result in the inequality $\mathbb{E} X(t \wedge \tau) \leq$ $\mathbb{E} X(t)$. When $\tau$ is a stopping time, this intuition is correct. Once we have Theorem 8.8.1 for submartingales, we easily obtain it for supermartingales by using the fact that the negative of a supermartingale is a submartingale. Since a martingale is both a submartingale and a supermartingale, we obtain the equality $\mathbb{E} X(t \wedge \tau)=\mathbb{E} X(t)$ for martingales.

Use Theorem 8.8.1 and Lemma 8.5.1 to show in the context of Subsection 8.5.1 that
$$
\widetilde{\mathbb{E}}\left[e^{-r T}(S(T)-K)^{+}\right]=\max _{\tau \in \mathcal{T}_{0, T}} \widetilde{\mathbb{E}}\left[e^{-r \tau}(S(\tau)-K)^{+}\right],
$$
where as usual we interpret $e^{-r \tau}(S(\tau)-K)^{+}$to be zero if $\tau=\infty$. The righthand side is the American call price analogous to Definition 8.4.1 for the American put price. The left-hand side is the European call price.

Check back soon!
01:59

Problem 7

A function $f(x)$ defined for $x \geq 0$ is said to be convex if, for every $0 \leq x_1 \leq x_2$ and every $0 \leq \lambda \leq 1$, the inequality
$$
f\left((1-\lambda) x_1+\lambda x_2\right) \leq(1-\lambda) f\left(x_1\right)+\lambda f\left(x_2\right)
$$
holds. Suppose $f(x)$ and $g(x)$ are convex functions defined for $x \geq 0$. Show that
$$
h(x)=\max \{f(x), g(x)\}
$$
is also convex.

Minh Le
Minh Le
Numerade Educator