Quantitative Finance
Final
Jul 7, 2023
Instruction
1. Download data
2. You might use calculators, R or Excel during the examination
3. The exam has been thought to be solved individually.
4. Your solutions has 1 Excel file, and your written solutions in paper.
5. You should upload your solutions to Blackboard and/or send it to [Link]@[Link],
make sure your solutions have been sent correctly.
1 Nonnormal Risk Metrics (6.0 points)
Assume that daily log-returns behave like
Rt+1 = σt zt
2
σt+1 = ω + αRt2 + βσt2
iid
where zt ∼ t(d) (standardized t-Student),
− 1+d
x2
2
ft(d) (x) = C(d) 1 +
d−2
and
Γ d+1
2
C(d) = p
Γ d2
π(d − 2)
1. (1.0 point) Obtain Et [Rt+1 ] and Vt (Rt+1 )
2. (2.0 point) Find the 1-day VaRpt+1 [Hint: You can express your solution in terms of the
standardized t-Student cdf, Ft(d) (x)]
3. (2.0 point) Find the 1-day ESpt+1 [Hint: Compute the conditional expectation. Remember
that conditional pdf is fX|X<h (x) = FfX (x)
X (h)
]
4. (1.0 point) Assume d = 5 and p = 0.01. Find the ratio between the normal 1-day VaRpt+1 to
nonnormal 1-day VaRpt+1 [Hint: Use R]
2 VaR and the ARMA(1,1) process (6.0 points)
Suddenly, you woke up in a universe where daily log-returns behave like
Rt+1 = µ + φ1 Rt + εt+1 + θεt
εt = σt zt
2
σt+1 = ω + αRt2 + βσt2
iid
where zt ∼ N (0, 1)
1
1. (1.0 point) Obtain Et [Rt+1 ] and Vt (Rt+1 )
2. (1.0 point) Find the 1-day VaRpt+1
3. (1.0 point) Find the 1-day ESpt+1
4. (2.0 point) After estimating µ, φ1 , θ, ω, α, β and given today’s portfolio value, VP F,t = 10
0.01
million. Compute the 1-day VaRt+1 and $VaR0.01
t+1 [Hint: Use the information in Excel file]
0.01
5. (1.0 point) Compute the 1-day ESt+1 and $ES0.01
t+1 [Hint: Use the information in Excel file]
3 Gaussian copula model of correlated default times (5.0
points)
In this exercise, you extend the Gaussian copula model to include two factors (market and macro
factors). I gave you and Excel with the information you need to compute the the Monte Carlo
estimation of Expected Loss, Discounted Expected Loss and Default Rate and also plot the distri-
bution of losses. Using 5000 Monte Carlo simulations
1. (2.0 points) Plot a histogram for discounted expected loss in a new sheet in two scenarios,
(a) without default dependence and (b) with default dependence
2. (2.0 points) Estimate the portfolio expected loss, discounted expected loss and default rate
using Monte Carlo with 5000 simulations in the scenario with default dependence
3. (1.0 points) Compute a 99% confidence interval for the previous estimation
4 CAPM in two periods (3.0 points)
Suppose the payoff from an investment in 2 years is X. What is the fair price, P , for this invest-
ment?
1. (0.5 points) Obtain the annual return on this investment.
2. (0.5 points) Obtain the βX of this investment as a function fo X, rm and P
3. (2.0 points) Suppose that CAPM holds, i.e. µX = rf + βX (µm − rf ). Find the fair price P
as a function of everything else.
