FIN500Q, Spring 2023 Quantitative Risk Management
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
FIN500Q, Spring 2023
Quantitative Risk Management
1. (10 points) Single Select Multiple Choice Questions. Write your answer in the box.
(a) (2 points) C
long words arises from the potential that a borrower or counterparty will fail to perform on an obligation.
A. Operational risk B. Market risk C. Credit risk
(b) (2 points) A
Which one of the following statements is correct?
A. Maximum likelihood methods are usually used to estimate parameters in GARCH(1,1) models from historical data.
B. For 2 random variables, independence is the same as 0 correlation.
C. An EWMA model gives equal weights to the square of historical returns.
(c) (2 points) A
The market portfolio consists of two stocks, stock 1 and stock 2, with respective betas 0.5 and -1.5. What are the weights of stock 1 and stock 2? (Hint: The market portfolio has a beta of 1.)
A. 1.25; -0.25 B. 0.75; 0.25
C. -1.25; 0.25 D. 0.25; 0.75
(d) (2 points) D
Which statement regarding Value at Risk (VaR) is correct?
A. VaR is always subadditive.
B. It is impossible to compute VaR for discrete distributions.
C. VaR is the same as sample standard deviation.
D. VaR is coherent for elliptical distributions.
(e) (2 points) B
An investor has a portfolio which consists of two stocks that have normally dis- tributed returns. Choose the correct statement about the VaR of the portfolio payoff compared to the VaRs of the individual stock positions.
A. The portfolio’s VaR is always greater than or equal to the sum of the individual VaRs of the two stocks.
B. The portfolio’s VaR is always smaller than or equal to the sum of the individual VaRs of the two stocks.
C. The portfolio’s VaR is always smaller than or equal to the sum of the individual VaRs of the two stocks only when the correlation is positive.
D. The portfolio’s VaR is always greater than or equal to the sum of the individual VaRs of the two stocks only when the correlation is negative.
2. (20 points) A firm produces a unit of gold a year from today. The gold price risk is diversifiable, and the price of gold next year is either $480 or $200, each with probability 0.5. Assume that the riskless rate is r = 6% and bankruptcy costs are $30 per unit of gold. The firm pays taxes at a rate of 40% on any cash owes in excess of $300. Interest payments on debt are tax deductible.
(a) (6 points) Suppose the firm hedges the gold price risk by entering into a forward
contract and chooses the optimal amount of debt. Then
. Would the firm choose to issue safe debt or risky debt, assuming that lenders
determine the yield on the debt?
. What is the firm value given that choice?
Answer:
The hedged firm has certain future cash flow, thus it can only issue safe debt. The future expected cash flow is 480 × 0.5 + 200 × 0.5 = $340.
The face value F of safe debt issued must satisfy
1.06F + 0.4(340 − 300 − 0.06F) ≤ 340
Equation says that the hedged firm’s debt payments(1.06F) and tax payments (0.4(340 − 300 − 0.06F)) equal the cash flow of $340.
Then we can solve for F = $312.74.
Therefore, the hedged firm’s tax payment is 0.4(340−300−0.06×312.74) = $8.494, and the firm value is
(340 − 8.494) = $312.74
Note that we should expect that the firm value is the same as the optimal face value of safe debt.
(b) (3 points) If the hedged firm issue $250 safe debt, then what is the firm value?
Answer:
The tax payment is 0.4(340 − 300 − 0.06 × 250) = $10.
Therefore, the firm value is
(340 − 10) = $311.32
Continued Question 1:
(c) (3 points) If the firm does not hedge the gold price risk and issues the optimal
amount of safe debt, then what is the firm value?
Answer:
The face value F of safe debt issued must satisfy
1.06F ≤ 200
Because in low cash flow state, there is no tax payment. Then the maximum F is 200/1.06 = $188.68
In high cash flow state, the tax payment is 0.4(480 − 300 − 0.06 × 188.68) = $67.47. Therefore, the expected tax payment is 0.5 × 0 + 0.5 × 67.47 = $33.74
The firm value is
(340 − 33.74) = $288.92
(d) (3 points) If the firm issues $250 of risky debt, find the yield on the risky debt and
the value of the unhedged firm.
Answer:
The yield on the risky debt, x, must be high enought that the expected return of the debtholders equals the riskless rate. Therefore,
0.5(200 − 30) + 0.5(250(1 + x)) = 250(1.06)
Then we can solve for x = 0.44.
In low cash flow state, there is no tax payment, the future cash flow is 200 − 30 = $170.
In high cash flow state, the tax payment is 0.4(480 − 300 − 250 × 0.44) = $28, the future cash flow is 480 − 28 = $452.
Therefore, the expected future cash flow is (0.5)(170) + (0.5)(452) = 311, the firm
value is
= $293.40
Continued Question 1:
(e) (5 points) Now suppose that the unhedged firm chooses the face value of risky debt
optimally. Find the face value of debt, the yield on the risky debt, and the value of the unhedged firm.
Answer:
The optimal face value F of debt issued and the corresponding yield on the debt x must satisfy the following two questions.
First, in high cash flow state, the future cash flow must cover all payments (debt payments and tax payments).
(1 + x)F + 0.4(480 − 300 − xF) = 480 (1)
Second, the yield on the debt,x, must be high enough that the expected return of the debtholders equals the riskless rate.
(2)
Based on equation (1) and (2), we can get the x and F .
x = 0.5627 F = 305.02
In low cash flow state, there is no tax payment, the future cash flow is 200 − 30 = $170.
In high cash flow state, the tax payment is 0.4(480−300−305.02×0.5627) = $3.35, the future cash flow is 480 − 3.35 = $476.65.
Therefore, the expected future cash flow is (0.5)(170) + (0.5)(476.65) = 323.325,
the firm value is
= $305.02
Note that we should expect that the firm value is the same as the optimal face value of debt.
3. (20 points) An investor has a $1,000 portfolio of two stocks TSLA and RIVN, each of which has a normal distribution of returns. The annual variance of the return of his portfolio is 0.04. The annual mean returns of the two stocks are 0.2 (TSLA) and 0.1 (RIVN). The betas of the two stocks with respect to his portfolio are 0.9 (TSLA) and
1.4 (RIVN). His portfolio weight in TSLA is 0.8.
(a) (5 points) Calculate the daily portfolio VaR(5%).
Answer:
First, the investor’s portfolio weight in RIVN is 1-0.8=0.2.
Second, the annual portfolio mean return is (0.2)(0.8) + (0.1)(0.2) = 0.18.
Third, the daily portfolio mean return is 0 .18/252 = 0.000714
Fourth, the daily portfolio standard deviation, σP =^(^) = 0.013
Fifth, use the formula to get the daily portfolio VaR(5%)
−|$1000| × (0.000714 + (−1.645) × 0.013) = $20.67
(b) (5 points) Using the marginal VaR(5%), find the approximate change in the 5%
daily portfolio VaR in dollars if the investor increases his position in TSLA by $200 dollars.
Answer:
Second, the daily TSLA mean return is 0.2/252 = 0.000794.
Third, the TSLA beta with respect to the portfolio is 0.9
Fourth, the marginal VaR of TSLA is
∆VaRα(i)(Rp ) = − [µi + zα · σRp · βi,Rp] = − [0.000794 + (−1.645)(0.013)(0.9)]
=0.01845
Fifth, the approximate change in the daily portfolio VaR(5%) in dollars is $200 × 0.01845 = $3.69
Continued Question 2:
(c) (10 points) A risk manager decides to backtest this 5% daily VaR model. Using a sample of 2,000 days, she produces the following statistics:
On 104 days, the loss was bigger than the daily VaR (a ”violation”);
On 1806 days, there was no violation the next day after no violation on that day; On 90 days, there was a violation the next day after no violation on that day; On 88 days, there was no violation the next day after a violation on that day; On 16 days, there was another violation the next day after a violation on that day. What is the value of the test statistic for independence of VaR violations?
Answer:
In this problem T = 2000,N = 104, therefore, π0 = 1 − = 0.948 and π 1 = = 0.052. Then N00 = 1806,N01 = 90,N10 = 88 and N11 = 16, thus
π00 = π01 = π 10 =
π 11 =
N00
N00 + N01 N01 |
N00 + N01 N10 |
N10 + N11 N11 |
N10 + N11
= 1806/(1806 + 90) = 0.9525
= 90/(1806 + 90) = 0.0475
= 88/(88 + 16) = 0.8462
= 16/(88 + 16) = 0.1538
We test the independence of VaR violations using the following formula
LRI =2 · ln [ ()N00 · ( )N01 · ( )N10 ( )N11]
=15.52
4. (20 points) Suppose that two variables V1 and V2 have exponential distributions with λ parameters of 2.9 and 2.3, respectively. The cumulative probability distribution for an exponential distribution is F(x) = 1 − e −λ北 . Use a Gaussian copula to define the correlation structure between V1 and V2 .
(a) (5 points) What is the probability that V1 ≤ 0.1 and what is the probability that
V2 ≤ 1?
Answer:
The probability that V1 ≤ 0.1 is
F(0.1) = 1 − e −2.9×0.1 = 0.2517
The probability that V2 ≤ 1 is
F(1) = 1 − e −2.3×1 = 0.8997
(b) (5 points) With a copula correlation of 0.5, what is the probability that V1 ≤ 0.1
and V2 ≤ 1?
Answer:
Using the information we get from previous questions, we can get the following results via the Table 1 on formula sheet.
Φ − 1 (0.25) = −0.674
Φ − 1 (0.90) = 1.282
Use Table 4(ρ = 0.5), we can get
P(V1 ≤ 0.1,V2 ≤ 1) = M(−0.67, 1.28, 0.5) = 24.74%
Continued Question 3:
(c) (10 points) Assume V1 and V2 are independent, that is, with a copula correlation
of 0, what is the probability that V1 > 0.1 and V2 > 1?
Answer:
Method 1 (Applied in any ρ):
P(V1 > 0.1,V2 > 1) =1 − P(V1 ≤ 0.1) − P(V2 ≤ 1) + P(V1 ≤ 0.1,V2 ≤ 1) =1 − 0.2517 − 0.8997 + 0.2262
=7.48%
where P(V1 ≤ 0.1,V2 ≤ 1) = 0.2262 can be obtained from Table 3(ρ = 0). Method 2:
P(V1 > 0.1) = 1 − F(0.1) = 1 − 1 + e −2.9×0.1 = 0.7483 P(V2 > 1) = 1 − F(1) = 1 − 1 + e −2.3×1 = 0.1003
Because V1 and V2 are independent,
P(V1 > 0.1,V2 > 1) = (0.7483)(0.1003) = 7.48%
Note that the defination of cumulative distribution function is M(x,y,ρ) = Prob(X ≤ x,Y ≤ y), but in the problem, we need to compute P(V1 > 0.1,V2 > 1).
Some of you first compute the Φ − 1 (0.7483) = 0.67 and Φ − 1 (0.1003) = −1.28, then directly claim that P(V1 > 0.1,V2 > 1) = M(0.67, −1.28, 0) = 7.51%.
This result is held only because the bivariate standard normal distribution is sym- metric.
2023-05-18