Exam 3 – Question 28

[Nitin Iyer]

Exam 3 – Question 28

[Anonymous]

The estimate of historical VaR at 99% confidence based on a set of data with 100 observations will end up being:
(a) the extrapolated returns of the last 1.64 observations
(b) the worst single observation in the data set
(c) the weighted average of the top 2.33 observations
(d) None of the above
riskprep says the correct answer is choice ‘b’

wouldnt the 99% VaR for 100 obs be the worst but one (and not the worst)?

so may be the answer is (d)

can someone pls explain how the answer is (b)?

[Anonymous]

I agree with (b). Maybe the confusion arises because we are dealing with a discrete rather than continuous situation. But in a continuous distribution we are looking for a value that exceeds 99% of possible values. With 100 observations only the worst observation can exceed 99 of the observations.

[Anonymous]

Hey i came across the below question on eduleap for exam iii… the correct answer is a).. i thought normal distribution was a necessary assumption in all simulations.. can anyone clarify.. is the answer wrong??

Cleared the 1st two exams..just used the mock tests.. any tips for exam iii would be very helpful.

Which of the following is not true for Monte Carlo simulations
A ) It has normal distribution of risk factors as an underlying assumption
B ) A large number of prices paths need to be simulated
C ) The distribution of the terminal values will depend on the process used to model the risk factor
D ) Jumps in underlying risk factors can be incorporated whilst simulating

[Anonymous]

Hi,

Answer A is the correct answer.
The shape of a distribution is a direct consequence of the math model that has been used to generate the values.
In MonteCarlo simulation, each risk factor can have its own model (FX will definitely be different from Commodity). Normal distribution is unlikely to be used in MonteCarlo simulation…as no Risk Factor runs with perfect Normal Distribution on the market.

[Author]

Posts