Prof. Chow Ying-Foon, Ph.D.

2021—2022 Second Term

Room 1215, Cheng Yu Tung Building Phone: 3943 7638
Department of Finance Fax: 2603 6586
The Chinese University of Hong Kong Email: [email protected]

FINA 4150 A Quantitative Methods for Financial Derivatives

Assignment 2
Due Date: February 22, 2022

Answer the following questions for a total of 200 points (questions are equally weighted)
and show all your work carefully. Your assignment will be checked for correctness, com-
pleteness, and clearness of the answers. I will also check your program (if applicable)
and run it myself. If necessary, I will ask you to give me a demo. In addition, I will
probably ask you further questions regarding your work. In this case, your response
will be also evaluated.
P P∞ 1 P∞
1. Show that the series ∞ =1 
1
2 is convergent, while the series =1  and =2
1
 ln()
are
divergent, i.e., equal to ∞. Note: It is known that
̰ !
X∞
1 2 X1
2
= and lim − ln() = 
=1
 6 →∞
=1


where  ≈ 057721 is called Euler’s constant. Hint: Note that

X
−1 Z  X
−1 Z +1 X1−1
1 1 1
 ln() =  =  
=1
+1 1  =1   =1


1 1 1
since +1
 
 
for any      + 1. Conclude that

1 X1

ln() +   ln() + 1 ∀ ≥ 1
 =1 
P
For the series ∞ 1
=2  ln() , use a similar method to find upper and lower bounds for the
1
integral of  ln() over the interval [2 ].

2. a. Many textbooks determine the formula for continuously compounded interest through
an argument which avoids the use of l’Hôpital’s rule.
¡ Consider
¢ the compound
 
amount  that satisfies the equation  = lim→∞  1 +  . Let  = . Then
³  ´
 1+ =  (1 + )(1)

and we can focus on finding the lim→0 (1 + )1 . Show that

(1 + )1 = (1) ln(1+)

and take the limit of both sides as  → 0. Hint: You can use the definition of the
derivative in the exponent on the right-hand side.

1
 b. Consider an investment of amount  now and receive a sequence of positive payoffs
{1  2       } at regular intervals. Use the Mean Value Theorem to show the
rate of return defined by the root of the following function is unique:
X

() = − +  (1 + )−
=1

Note: We can see that  () is continuous on the open interval (−1 ∞). In the limit
as  approaches −1 from the right, the function values approach positive infinity.
On the other hand as  approaches positive infinity, the function values approach
−  0 asymptotically. Thus by the Intermediate Value Theorem there exists ∗
with −1  ∗  ∞ such that (∗ ) = 0.
3. a. Let
1 − 2
( ) = √  4 for   0  ∈ R
4
 2
Compute 
and 2
, and show that
  2 
= 2
 
Note: This exercise shows that the function ( ) is a solution of the heat equation.
In fact, ( ) is the fundamental solution of the heat equation, and is used in the
PDE derivation of the Black—Scholes—Merton formula for pricing European plain
vanilla options. Also, note that ( ) is the same as the density function of a
normal variable with mean 0 and variance 2.
b. Consider the Gamma of a plain vanilla European option in Black—Scholes—Merton
world: µ 2¶
(1 ) 1 
Γ= √ = √ exp − 1
   2 2
where ¡¢ ³ ´
2
ln  +  + 2 
1 = √
 
Show that, as a function of   0, the Gamma is first increasing until it reaches a
maximum point and then decreases. Also, show that
lim Γ() = 0 and lim Γ() = 0
→0 →∞

4. a. Find the maximum and minimum of the function  (1  2  3 ) = 42 − 23 subject
to the constraints 21 − 2 − 3 = 0 and 21 + 22 = 13.
b. Assume that you can trade for four assets (and that it is also possible to short the
assets). The expected values, standard deviations, and correlations of the rates of
return of the assets are:
1 = 008 ; 1 = 025 ; 12 = −025
2 = 012 ; 2 = 025 ; 23 = −025
3 = 016 ; 3 = 030 ; 13 = 025
4 = 005 ; 4 = 020 ; 4 = 0 ∀ = 1 : 3

2
 i. Find the asset allocation for a minimal variance portfolio with 12% expected
rate of return.
ii. Find the asset allocation for a maximum expected return portfolio with stan-
dard deviation of the rate of return equal to 24%.
5. a. Let ,  ,  and  be positive constants, and define the function  : R → R as
Z ()
1 2
() = √ − 2 
2 0
³ ¡ ¢ ³ 2
´ ´ ³ √ ´
where () = ln  +  + 2     . Compute 0 (). Note: This func-
tion is related to the Delta of a plain vanilla call option.
R∞
b. Let () be a function such that −∞ | ()|  exists. Define () by
Z ∞
() = ( − )()


and show that

2
= ()
2
Note: The price of a call option can be regarded as a function of the strike price
. By using risk-neutral valuation, we can show that
Z ∞
− b −
() =  E(max(( ) −  0)) =  max( −  0) ()
−∞
Z ∞ Z ∞
−
=  ( − ) () = ( − )()
 

where  () is the probability density function of ( ) given (0), and () =
− (). Then, according to the result of this exercise,
2
= −  ()
 2
6. a. Compute the integral of the function ( ) = 2 − 2 on the region bounded by
the parabola  = ( + 1)2 and the line  = 5 − 1.
R3√
b. Compute an approximate value of 1 −  using the Midpoint rule, the Trape-
zoidal rule, and Simpson’s rule. Start with  = 4 intervals, and double the number
of intervals until two consecutive approximations are within 10−6 of each other.
7. a. Let  be the domain bounded by the -axis, the -axis, and the line  +  = 1.
Compute Z Z
−
 
 +
Hint: Use the change of variables  =  +  and  =  − , which is equivalent to
 = +
2
and  = −
2
. Note that ( ) ∈  if and only if 0 ≤  ≤ 1 and − ≤  ≤ .
b. Use the change of variables to polar coordinates to show that the area of a circle
of radius  is 2 , i.e., prove that
Z Z
1   = 2
(0)

3
 8. Part of a well-known puzzle involves three people entering a room. As each person enters,
at random either a red or a blue hat is placed on the person’s head. The probability that
an individual receives a red hat is 1/2. No person can see the color of their own hat,
but they can see the color of the other two persons’ hats. The three will split a prize
if at least one person guesses the color of their own hat correctly and no one guesses
incorrectly. A person may decide to pass rather than to guess. The three people are
not allowed to confer with one another once the hats have been placed on their heads,
but they are allowed to agree on a strategy prior to entering the room. At the risk of
spoiling the puzzle, one strategy the players may follow instructs a player to pass if they
see the other two persons wearing mis-matched hats and to guess the opposite color if
their friends are wearing matching hats. Why is this a good strategy and what is the
probability of winning the game?

9. Suppose we have a coin that when tossed results in heads with probability , and tails
with probability 1 − , where  ∈ (0 1). Let  be the number of times that we need to
toss the coin (denoted ) until  heads are obtained. Obviously,  ≥  ≥ 1.

a. What is Pr( = )? Show this is indeed a probability mass function.

b. Compute E().
c. Compute Var().
d. What is Pr( = |1-st head occurs on the 5-th toss)?

10. Suppose X is a Gaussian random variable with mean μ and covariance matrix Σ, in 
dimensions.

a. Let B be an  ×  real matrix. The scalar random variable  = X0 BX is referred

to as a quadratic form (in normal variables). Show that if B is not symmetric,
its coefficients can be arranged into  = X0 AX where A is an  ×  symmetric
matrix.
b. Find E(X0 AX)
0
c. E(X AX )

11. You have 10 independent coins, each having probability 35 of flipping heads. Let  be
the number of heads flipped when all ten coins are flipped.

a. What is Pr( = 6)?

b. What is Pr( ≥ 6)?
c. What is E()?
d. What is Var()?

4
12. a. You have a deck of 52 cards in which there are 4 cards of each denomination
1 2     13.
i. In how many ways can you draw a full house, which is a hand of 5 cards with
denominations ?
ii. What is the probability of obtaining a full house?
b. In how many ways can you sample 10,000 balls from a set of 100,000 distinct balls.
Give a decent estimate, not just ∞. (You may either numerically try to estimate
this by (say) first estimating the logarithm of this number, or you may use an
analytic approximation to the factorial function.)

13. You roll two (fair) dice.

a. What is the probability that at least one of the dice has rolled a 4?
b. You are told the sum is . Conditioned on the value of , for  = 2 3     12,
what is the probability that at least one of the dice has rolled a 4 now?
c. Compute the probability that ‘the sum is 10’ conditioned on ‘at least one of the
dice has rolled a 4’.

14. Let  be a positive integer with 2 ≤  ≤ 12. You throw two fair dice. If the sum of the
dice is , you win (), or lose 1 otherwise. Find the smallest value of () that makes
the game worth playing.

15. a. Let  be a non-negative random variable with mean . For   0, show that

P( ≥ ) ≤


b. Let  be a random variable with mean  and variance  2 . Then for   0, show
that
2
P(| − | ≥ ) ≤ 2

16. a. If  has a uniform
√ distribution on [0 1], what are the probability density functions
of ln  and ?
b. Let 1       be independent exponential random variables having density () =
− . Let 1       be random variables defined by

1 = 1
2 = 1 + 2
..
.
 = 1 + 2 + · · · + 

i. Find the density of  , using induction.

ii. Now do the same using moment generating functions.
iii. How is the distribution related to the Gamma density?

5
 P
17. a. You have numbers 1       . The mean is  = 1  =1  and the variance is
2 1
P 2
  =  =1 ( − ) .
P
i. Show that 2 = 1  2 2
=1  −  .
ii. Show that for any set of numbers, the average of the squares is as large as the
square of the average.
iii. How can you compute the variance with just one pass through the data 1       .
b. You have a stream of independent uniform random variables taking on values in
[0 1].
i. How will you generate a uniform random variable in the interval [ ]?
ii. How will you generate a Gaussian random variable with mean 0 and variance
1?
iii. How will you generate a Gaussian random vector x with mean vector μ and
covariance matrix Σ?

18. The density function of the exponential random variable  with parameter   0 is
½ −
 if  ≥ 0
 () =
0 if   0

a. Show that the function () is indeed a density function. It is clear that () ≥ 0,
for any  ∈ R. Prove that Z ∞
() = 1
−∞

b. Show that the expected value and the variance of the exponential random variable
 are E() = 1 and Var() = 12 .
c. Show that the cumulative density function of  is
½
1 − − if  ≥ 0
 () =
0 otherwise

d. Show that Z ∞
P( ≥ ) = () = −

Note: this result is used to show that the exponential variable is memoryless, i.e.,
P( ≥  + | ≥ ) = P( ≥ ).

6
19. a. Suppose the continuous random variable  has the probability density function
½
(42 + 6)  ∈ [0 2]
 () =
0 otherwise

for some constant .

i. What is E()?
ii. Calculate Pr(  15).
b. Suppose the continuous random vector (  ) has the joint probability distribution
½
(42  +  2 )  ∈ [0 1]  ∈ [0 1]
( ) =
0 otherwise

for some constant .

i. Calculate Pr( +   15).
ii. Calculate E( ).
iii. What is Cov(  )?
iv. What is  (|), the conditional probability density function of  given ?
v. Compute E(| ) = 05.

20. The conditional expectation identity states that for random variables,  and  , we have

E() = E(E(| ))

Similarly, the conditional variance identity states that

Var() = Var(E(| )) + E(Var(| ))

now let  = 1 + · · · +  where the  ’s are i.i.d. and  is also a random variable
independent of the  ’s.

a. Use the conditional expectation identity to show that E( ) = E(1 ) E().
b. Use the conditional variance identity to show that Var( ) = E(1 )2 Var() +
Var(1 ) E().