MGFD30: Risk Management

Sample Test

Instructor: A. Mazaheri

Instructions: This is a closed book examination. You are permitted to bring a non-programmable
calculator and a hand-written one side of one 8½” × 11” page with notes and/or formulae. Show all your
work otherwise you will not get full credit.

Make sure you allocate time appropriately.

You have 2 hours.

Good Luck!

NAME: _____________________________________________

ID#: _____________________________________________

Problem-1 _______________ (30 points)

Problem-2 _______________ (15 points)

Problem-3 _______________ (15 points)

Problem-4 _______________ (15 points)

Total _____________ (75 points)

Page 1 of 21
Problem-1 [20 Points]: Answer the following short questions.

a) [5 Points] An investment banking institution uses the GARCH(1,1) model to estimate the
volatility. The three parameters of this model are ω, α and β. Briefly explain the impact of
making a small increase in each of the parameters while keeping the others fixed.

b) [5 Points] It has been argued that the distribution of the exchange rate returns exhibits heavy
tails. Using a graph explain what this means?

Page 2 of 21
c) [5 Points] Why is there an add-on amount in Basel I for derivatives transactions? “Basel I
could be improved if the add-on amount for a derivatives transaction depended on the value of
the transaction.” How would you argue this viewpoint?

Page 3 of 21
d) (5 Points) The following table provides the add-on factors for derivatives under Basil I:

Remaining Interest rate Exch Rate Equity Precious Metals Other

Maturity (yrs) and Gold except gold Commodities

<1 0.0 1.0 6.0 7.0 10.0

1 to 5 0.5 5.0 8.0 7.0 12.0

>5 1.5 7.5 10.0 6.0 15.0

An OECD bank has the following transactions with another OECD bank:

I) A long position in a six-month option on the S&P 500. The principal is $20 million and the
current value is $6 million.
II) A two-year swap involving oil. The principal is $30 million and the current value of the swap
is –$5 million.

Assuming a risk weight of 20%, what is the capital requirement for this bank without and with
netting?

Page 4 of 21
e) (5 Points) A risk manager assumes that the joint distribution of returns is multivariate normal
and calculates the following risk measures for a 2-asset portfolio:

Asset Position Individual VaR Marginal VaR Incremental VaR

1 $100 $23.3 0.176 $17.6
2 $100 $46.6 0.440 $44.0
Portfolio $200 $61.6 $61.6

Let βi = ρip × σi / σp, where ρip denotes the correlation between the return of asset i and the return
of the portfolio, σi is the volatility of the return of asset i and σp is the volatility of the return of
the portfolio. What is β1?

f) (5 Points) An investor has a portfolio that includes three assets with the following information:

1 2 1 2
𝜇𝜇 = �2� , 𝐶𝐶 = �1 5 3�
3 2 3 6
0.1
The weight vector of your portfolio is: 𝛼𝛼 = �0.3�. The investor has invested $100,000 in this
0.6
portfolio. Compute the 1-Day 99% VaR for her portfolio.

Page 5 of 21
Problem-2: [15 points] The following GARCH(1,1) model is used to forecast the daily return
variance of an asset:
2 2
𝜎𝜎𝑛𝑛2 = 0.00004 + 0.08𝑢𝑢𝑛𝑛−1 + 0.82𝜎𝜎𝑛𝑛−1

a) (5 Points) Suppose the estimate of the volatility today is 6.0% and the asset return is -4.0%.
What is the estimate of the long-run average volatility per day? What is the estimate of the new
volatility?

b) (5 Points) If the current (annualized) volatility is 25% per year, find the expected volatility in
20 days.

Page 6 of 21
c) (5 Points) The GARCH models are estimated using MLE method. Explain how this
estimation method works. To get full mark you need to start by explaining what a likelihood
function is.

Page 7 of 21
Question-3 [15 Points]: A trader has a book of portfolio consisting of some stocks of a single company
paying a continuous dividend, some call options, and some put options. Both options are in European
style and defined on the same stock. Suppose the current stock price S = $60, time to maturity 0.25 year,
the annual interest rate r = 0.062, the dividend-yield is 2%, the annual volatility 0.4, and the strike prices
for call and put are $70 and $50 respectively.

a) [5 Points] Fill in the following Table:

Type Delta Gamma Vega

S (Stock)
P (Put)
C (Call)

Page 8 of 21
b) [5 Points] The trader is holding 1000 units of S, short 200 units of C and long 100 units of P.
What are the total Delta, Gamma, and Vega of this portfolio?

c) [5 Points] The trader wishes to hedge this portfolio with two call options given the following
characteristics

Type Delta Gamma Vega

C1 0.5 0.03 10
C2 0.6 0.02 8

Suppose the trader wants to create a portfolio that is Gamma and Vega neutral. Could he/she
accomplish this objective using two options, C1 and C2?

Page 9 of 21
Question-4 [10 Points]: A stock with a price of 60 and annual volatility of 0.4 has a call option with a
delta of 0.5 and gamma of 0.2 and a put with delta of -0.5 and gamma of 0.1.

a) (5 Points) What is 10-day 99% VaR for the portfolio of 1000 stocks?

b) (5 Points) Based on the linear model, what is 10-day 99% VaR for the portfolio of 1000 stocks, short
200 call options, and long 100 put options?

Page 10 of 21
c) (5 Points) What is the shape of distribution of your portfolio in part (b), normal, right skewed, left
skewed? Why?

Page 11 of 21
Page 12 of 21
 Solutions
Problem-1 [20 Points]: Answer the following short questions.

a) [5 Points] An investment banking institution uses the GARCH(1,1) model to estimate the
volatility. The three parameters of this model are ω, α and β. Briefly explain the impact of
making a small increase in each of the parameters while keeping the others fixed.

Answer:

ω: As we know the weight given to the long-run average variance rate is 1 − α − β and the
long-run average variance rate is ω/(1 − α − β). Therefore, an increasing in parameter ω will
increase the long-run average variance rate.

α: Increasing this parameter increases the weight given to the most recent data item and reduces
the weight given to the long-run average variance rate. This will also increase the level of the
long-run average variance rate.

β: Increasing this parameter increases the weight given to the previous variance estimate and
reduces the weight given to the long-run average variance rate. This will also increases the
level of the long-run average variance rate.

b) [5 Points] It has been argued that the distribution of the exchange rate returns exhibits heavy
tail. Using a graph explain what this means?

Answer:
Exchange rates are said to exhibits heavy tail distribution in comparison to the normal
distribution (as you can see from the graph) because the big absolute changes are a lot more
pronounced than the normal distribution suggests. This is since the exchange rate market is
characterized by stability (not much change) and violent absolute changes which the normal
distribution cannot pick up.

Note: To be more precise the distribution of equity market (S&P500) is said to be heavy tail as in
the figure, while the exchange rates are negatively skewed (only the left tail is heavy)

Page 13 of 21
c) [5 Points] Why is there an add-on amount in Basel I for derivatives transactions? “Basel I
could be improved if the add-on amount for a derivatives transaction depended on the value of
the transaction.” How would you argue this viewpoint?

Answer:

The add-on amount is to allow for a possibility that the exposure will increase prior to a default.

To argue for a relationship between the add-on amount and the value of the transaction, consider
two cases:

1. The value of the transaction is zero.

2. The value of the transaction is –$10 million

The current exposure is zero in both cases. In the first case any increase in the value of the
transaction will lead to an exposure. In the second case the transaction has to increase in value by
more than $10 million before there is an exposure—and it might be very unlikely that this will
happen. However, the capital required is the same in both cases.

Page 14 of 21
d) (5 Points) The following table provides the add-on factors for derivatives under Basil I:

Remaining Interest rate Exch Rate Equity Precious Metals Other

Maturity (yrs) and Gold except gold Commodities

<1 0.0 1.0 6.0 7.0 10.0

1 to 5 0.5 5.0 8.0 7.0 12.0

>5 1.5 7.5 10.0 6.0 15.0

An OECD bank has the following transactions with another OECD bank:

I) A long position in a six-month option on the S&P 500. The principal is $20 million and the
current value is $6 million.
II) A two-year swap involving oil. The principal is $30 million and the current value of the swap
is –$5 million.

Assuming a risk weight of 20%, what is the capital requirement for this bank without and with
netting?

Answer:

I) 6 + 0.06 × 20 = 7.2
II) 0.12 × 30 = 3.6

Total credit equivalent amount = 7.2+3.6 = 10.8

The risk weighted amount is 10.8 × 0.2 = 2.16
The capital required is 0.08 × 2.16 = $0.1728 million.

With netting:
The current exposure is: 6 − 5 =1.

The NRR is therefore 1/6 = 0.1667.

The credit equivalent amount is in millions of dollars:

1 + (0.4 + 0.6 × 0.1667) × (0.06 × 20 + 0.12 × 30) = 3.4

The risk weighted amount is 0.2×3.4 = 0.68 and the capital required is 0.08×0.68 = 0.0544.

Page 15 of 21
e) (5 Points) A risk manager assumes that the joint distribution of returns is multivariate normal
and calculates the following risk measures for a 2-asset portfolio:

Asset Position Individual VaR Marginal VaR Incremental VaR

1 $100 $23.3 0.176 $17.6
2 $100 $46.6 0.440 $44.0
Portfolio $200 $61.6 $61.6

Let βi = ρip × σi / σp, where ρip denotes the correlation between the return of asset i and the return
of the portfolio, σi is the volatility of the return of asset i and σp is the volatility of the return of
the portfolio. What is β1?

Answer:

Marginal VaRi = βi×Portfolio VaR / Portfolio Value

So, βi = Marginal VaRi×Portfolio Value / Portfolio VaR

β1 = 0.176 × 200 / 61.6 = 0.5714

Page 16 of 21
f) (5 Points) An investor has a portfolio that includes three assets with the following information:

1
𝜇𝜇 = �2�
3
2 1 2
𝐶𝐶 = Σ = �1 5 3�
2 3 6
0.1
The weight vector of your portfolio is: 𝛼𝛼 = �0.3�. The investor has invested $100,000 in this
0.6
portfolio. Compute the 1-Day 99% VaR for her portfolio.

Answer: With the normality assumption, we know the exact portfolio distribution, which is also
normal with mean and variance:

𝜇𝜇𝑝𝑝 = 0.1 × 1 + 0.3 × 2 + 0.6 × 3 = 2.5

𝛼𝛼 𝑇𝑇 𝐶𝐶𝐶𝐶 = 4.01 = 𝜎𝜎𝑝𝑝2

Thus:

𝑉𝑉𝑉𝑉𝑉𝑉99% = �2.5 − 2.33 × √4.01� = |−2.1658|

2.5 1
𝑉𝑉𝑉𝑉𝑉𝑉99%,10 𝐷𝐷𝐷𝐷𝐷𝐷 = � − 2.33 × � √4.01� = |−0.28707|
365 252

If 100,000 invested the VaR will be:

𝑉𝑉𝑉𝑉𝑉𝑉99%,1 𝐷𝐷𝐷𝐷𝐷𝐷 = 0.28707% × 100,000 = 287.07

Page 17 of 21
Problem-2: [15 points] The following GARCH(1,1) model is used to forecast the daily return
variance of an asset:
2 2
𝜎𝜎𝑛𝑛2 = 0.00004 + 0.08𝑢𝑢𝑛𝑛−1 + 0.82𝜎𝜎𝑛𝑛−1

a) (5 Points) Suppose the estimate of the volatility today is 6.0% and the asset return is -4.0%.
What is the estimate of the long-run average volatility per day? What is the estimate of the new
volatility?
b) (5 Points) If the current (annualized) volatility is 25% per year, find the expected volatility in
20 days.
c) (5 Points) The GARCH models are estimated using MLE method. Explain how this
estimation method works. To get full mark you need to start by explaining what a likelihood
function is.

Solution:

a) The model corresponds to α = 0.08, β = 0.82, and ω = 0.00004. Because γ = 1− α− β, it

follows that γ = 0.10. Because the long-run average variance, VL, can be found by VL = ω / γ, it
follows that VL = 0.0004. In other words, the long-run average volatility per day implied by the
model is sqrt(0.0004) = 2%.

The new variance is:

2 2
𝜎𝜎𝑛𝑛2 = 0.00004 + 0.08𝑢𝑢𝑛𝑛−1 + 0.82𝜎𝜎𝑛𝑛−1

𝜎𝜎𝑛𝑛2 = 0.00004 + 0.08 × (−0.04)2 + 0.82 × 0.062 = 0.00312

So the estimate of volatility is 0.055857.

b)

The equation describing the way the variance rate reverts to its long-run average is
2 ]
𝐸𝐸[𝜎𝜎𝑛𝑛+𝑘𝑘 = 𝑉𝑉𝐿𝐿 + (𝛼𝛼 + 𝛽𝛽)𝑘𝑘 (𝜎𝜎𝑛𝑛2 − 𝑉𝑉𝐿𝐿 )
2
[𝜎𝜎𝑛𝑛+𝑘𝑘 ] = 0.0004 + (0.90)𝑘𝑘 (𝜎𝜎𝑛𝑛2 − 0.0004)

If the current volatility annualized is 25%, then, σn = 0.25/√252 = 0.015749. Therefore:

2
[𝜎𝜎𝑛𝑛+20 ] = 0.0004 + (0.90)20 (0.0157492 − 0.0004) = 0.0003813

Therefore, the expected variance in 20 days is √0.000248 = 0.019533 (1.9533%) per day.

c) The MLE estimation chooses the parameters of the variance ω, α, β such that the likelihood of the
existing sample is maximized. The likelihood function is a function that assigns probability to
the existing sample - assuming independent samples, it is ∏𝑛𝑛𝑖𝑖=1 𝑃𝑃𝑖𝑖 where P is the probability of an
a specific observation. In a typical MLE estimate of GARCH model the probability is assumed to
follow a normal distribution, and the corresponding likelihood function is maximized by
choosing the parameters.

Page 18 of 21
Question-3 [15 Points]: A trader has a book of portfolio consisting of some stocks of a single company
paying a continuous dividend, some call options, and some put options. Both options are in European
style and defined on the same stock. Suppose the current stock price S = $60, time to maturity 0.25 year,
the annual interest rate r = 0.062, the dividend-yield is 2%, the annual volatility 40%, and the strike prices
for call and put are $70 and $50 respectively.

a) [5 Points] Fill in the following Table:

S* = 60×e-0.02×0.25 = 59.70, r = 6.2%, σ = 40%, T = 0.25

For the call since K = 70, we have:

ln(59.70 / 70) + (0.062 + 0.42 / 2)(1 / 4)

d1 = = −0.6183
0.4 1 / 4

Delta = N(-0.618)× e-0.02×0.25 = 0.2669

 1  1 − 2d1 
2

Γ =   
σ  2Π e 
 0
S T  
0.6182
1 1
Gamma = [� � 𝑒𝑒 − 2 ]𝑒𝑒 −0.02×0.25 = 0.027462
59.7×0.4×0.5 √2Π

Vega = 59.72 × 0.25 × 0.4 × 0.0275 = 9.788

For the put since K = 50, we have:

ln(59.7 / 50) + (0.062 + 0.42 / 2)(1 / 4)

d1 = = 1.064
0.4 1 / 4

Delta = (N(1.064)-1)×e-0.02×0.25 = -0.148

1.0642
1 1
Gamma = �� � 𝑒𝑒 − 2 � 𝑒𝑒 −0.02×0.25 = 0.0189
59.7×0.4×0.5 √2Π

Vega = 59.72 × 0.25 × 0.4 × 0.0189 = 6.736

Therefore:

Type Delta Gamma Vega

S (Stock) 1 0 0
P (Put) -0.148 0.0189 6.736
C (Call) 0.266 0.0275 9.788

Page 19 of 21
b) [5 Points] The trader is holding 1000 units of S, short 200 units of C and long 100 units of P. What are
the total Delta, Gamma, and Vega of this portfolio?

Solution:

∆ Π = 1000 − 200∆ C + 100∆ P ≅ 932

ΓΠ = −200ΓC + 100ΓP = −3.61
ν Π = −200ν C + 100ν P = −1284

c) [5 Points] The trader wishes to hedge this portfolio with two call options given the following
characteristics

Type Delta Gamma Vega

C1 0.5 0.03 10
C2 0.6 0.02 8

Suppose the trader wants to create a portfolio that is Gamma and Vega neutral. Could he/she accomplish
this objective using two options, C1 and C2?

Solution:

ΓΠ = −200ΓC + 100ΓP = −3.61

ν Π = −200ν C + 100ν P = −1284
W1 × 0.03 + W2 × 0.02 = 3.61
W1 × 10 + W2 × 8 = 1284
W1 = 80
W2 = 60.5

Buy 80 of C1 and 60.5 of C2. Then calculate the delta of your portfolio and delta hedge by buying/selling
stock.

Page 20 of 21
Question-4 [10 Points]: A stock with a price of 60 and annual volatility of 0.4 has a call option with a
delta of 0.5 and gamma of 0.2 and a put with delta of -0.5 and gamma of 0.1.

a) (5 Points) What is 10-day 99% VaR for the portfolio of 1000 stocks?

Solution:
0.4
σ Daily = 60,000 × = 1,511.86
252
Var99% = 2.33 × σ Daily × 10 = 2.33 × 1,511.86 × 10 = 11,139.53

b) (5 Points) Based on the linear model, what is 10-day 99% VaR for the portfolio of 1000 stocks, short
200 call options, and long 100 put options?

Solution:

ΛP = ∑ S × δ × ∆xi = 1000 × 60 × ∆x + 60 × −200 × 0.5 × ∆x + 60 × 100 × (− 0.5) × ∆x = 51,000∆x

σ ∆p = 51,000 × 0.4 = 20,400
10
VaR.99 = 2.33 × 20,400 × = 9468.6
252

c) (5 Points) What is the shape of distribution of your portfolio in part (b), normal, right skewed, left
skewed? Why?

Solution:

The gamma of the portfolio is: -200×0.2 + 100×0.1=-30

Therefore the third moment will be negative => The distribution will be negatively (left) skewed.

Page 21 of 21