CQF Module 2 Examination: Instructions

CQF Module 2 Examination
June 2013 Cohort
Instructions
All questions must be attempted. Complete workings must be provided to obtain maximum credit.
Each plot must have a brief explanation.
Books and lecture notes may be referred to. Help from others is not permitted. If you are unclear
about any question or what is required, contact the tutor only on [email protected].
A. Strategies and Pricing [36%]

1. A European put option is traded at $4 with the underlying price at $95. Time to expiry is one
month and the strike is $100. The risk-free interest rate is 3% per annum. Devise a trading strategy
to explore the arbitrage opportunity. What is the minimum profit of the strategy?
2. To hedge a Binary Call option B, a trader creates a call spread by short-selling a vanilla call option
with the lower strike (below Binary’s) λ1 < 0 and buying a call option with the higher strike λ2 > 0.
B + λ1 C1 + λ2 C2 = 0
(a) Construct a payoff diagram for a call spread that offsets a Binary Call with the strike $100.
Note: use λ1 = −1 and λ2 = 1 to see the shape of the payoff.
(b) For the option prices listed in the table below, find symmetric values |λ1 | = |λ2 | that offset a
payoff of exactly one Binary Call quoted in the market at B = $0.5.
Option Strike Maturity Price

C1 $90 1 $14.81
C2 $110 1 $4.94
Construct a payoff diagram for a basket of all three options using the updated values of λ1 , λ2 .
3. Implement the multi-step Binomial Method to price a European put with the following parameters:
strike K = 100 and maturity T = 1. Asset price level S0 = 100 and interest rate r = 0.05.
(a) For the constant number of time steps in the tree NTS=4, calculate the value of the option
for a range of volatilities and plot the result.
(b) Then, fix volatility at σ = 0.2 and plot the value of the option as a function of the number of
time steps in the tree, NTS=1, 2, . . . , 50. You will need a different tree for each NTS value.
Note: This is a computational task. Preferred solution method is a function written in VBA but
Excel spreadsheets with binomial trees will be accepted given the plots are correct. For simplicity,
use compound rate as a substitute to simple (discrete) rate.
1
B. Portfolio Optimisation and Risk [44%]
Note: Part of this task is computational. It is best to present matrix manipulations on a spreadsheet.
You can use the necessary Excel functions such as MMULT(), MINV() and TRANSPOSE(). Alterna-
tively, you can use MATLAB or R if familiar with those environments.
Consider an investment universe composed of the following assets:
Asset µ σ
A 0.04 0.07
B 0.08 0.12
C 0.12 0.18
D 0.15 0.26
with a correlation structure  

1 0.2 0.5 0.3
 
 0.2 1 0.7 0.4 
R=
 0.5 0.7 1

 0.9 

0.3 0.4 0.9 1
Denote the column vector of asset weights by w, the column vector of asset returns by µ, and the
covariance matrix by Σ.
1. Compute the covariance matrix Σ = SRS, where S is a diagonal matrix of standard deviations.
2. Consider the following optimization task

1
min wT Σw
w 2
subject to constraints
wT 1 = 1
T
w µ = 0.1
(a) Explain in plain English the purpose of the optimization.

(b) Solve this optimization task analytically – provide workings for the Lagrangian method.
(c) Compute the standard deviation σΠ of this optimal portfolio allocation w∗ .
(d) Simulate the Efficient Frontier for this investment universe and identify this portfolio on it .
The number of points to simulate the Frontier is your choice.
3. Use the optimal allocations w∗ to estimate the Analytical VaR for the portfolio with c = 99%
confidence
√
VaR(X) = wT µ + wT Σw × Factor
where for the Factor is drawn from
(a) the Normal distribution Factor = Φ−1 (1 − c).

(b) the Student’s t distribution Factor = Tν−1 (1 − c) with ν = 30.
Student’s t distribution allows more realistic modelling of ‘fatter tails’ of returns distribution.
The distribution changes its shape depending on the degrees of freedom parameter ν. You can
use Excel function TINV() but note that it provides an answer for the two-tail Factor.
2
4. Calculate contributions of individual assets to Value at Risk, estimated with Student’s t Factor,
using wT VaR(X). Construct and bar chart.
5. Download the spreadsheet built during the Value at Risk lecture. For the column with the Estimated
P&L result, find the numerical answer to the Value at Risk by the following method:
(a) Sort the column in an ascending order. Create a parallel column with an index number from
1 to the N , i.e., i = 1, 2, ..., 1009
i
(b) represents the cumulative probability of each P&L figure Pr(X ≤ Sorted P&L). Use this
N
property to find the first percentile (i.e., 99% VaR figure).
(c) Check your result using Excel’s PERCENTILE() function.
C. Useful Processes and Martingales [20%]

A diffusion process Y (t) is a martingale if its SDE does not have a drift term. The SDE can be
constructed by evaluating partial derivatives of a function F (t, X) and substituting into
1 ∂2F

∂F ∂F
dY (t) = + dt + dX(t)
∂t 2 ∂X 2 ∂X
Assume that all Y (t) satisfy the integrability condition.
1 2
1. (a) Construct an SDE for the process Y (t) = eσX(t)− 2 σ t
and (b) show that the process is, in fact,
an Exponential Martingale.
Note: if the generic Exponential Martingale follows dZ(t) = Z(t)g(t)dX(t) as defined in Martin-
gales Fundamentals lecture, to show (b) identify the g(t) and Z(t) for this expression Y (t).
p
2. Y (t) = wX1 (t) + 1 − w2 X2 (t) is an expression to create a correlated BM from two uncorre-
lated BMs, where w is factor loading that links to correlation. Assume that X1 (t) and X2 (t) are
independent Brownian Motions with drifts, such that
dX1 (t) = µ1 dt + σ1 dW1 (t)

dX2 (t) = µ2 dt + σ2 dW2 (t)
Construct an SDE for Y (t) and answer the following questions:
(a) Is the process Y (t) a martingale?

(b) How does answer to (a) depends on the value of constant w? Consider the drift term.
(c) Will Y (t) keep the properties of the Brownian Motion? Consider the increment Y (t) − Y (s).
3. A Poisson random variable N (t) ∼ P o(λ) creates a jump process with the fixed increment ∆. The
probability of the jump between times s and t is Pr(∆) ≈ λ(t − s).
Now, consider a compensated Poisson process M (t) = N (t) − λt where E [N (t)] = λt. Use this
property in order to provide a discrete-time proof that M (t) is a martingale.
Note: to explain both conditions,
E [|M (t)|] < ∞ (the process is bounded)

E [M (t)|Fs ] = M (s) ∀t > s

CQF Module 2 Examination: Instructions

Uploaded by

Copyright:

Available Formats

CQF Module 2 Examination: Instructions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CQF Module 2 Examination: Instructions

Uploaded by

Copyright:

Available Formats

CQF Module 2 Examination

June 2013 Cohort

A. Strategies and Pricing [36%]

Option Strike Maturity Price

Consider an investment universe composed of the following assets:

with a correlation structure  

2. Consider the following optimization task

(a) Explain in plain English the purpose of the optimization.

(a) the Normal distribution Factor = Φ−1 (1 − c).

C. Useful Processes and Martingales [20%]

dX1 (t) = µ1 dt + σ1 dW1 (t)

Construct an SDE for Y (t) and answer the following questions:

(a) Is the process Y (t) a martingale?

E [|M (t)|] < ∞ (the process is bounded)

You might also like