- Thread starter Nicole Seaman
- Start date
- Tags errors revisepdf

Hi in R38 P2 T5 , Reading material , page 38 in spearman's Rank correlation , example has been taken from GARP reading, it would be better if table 8.1 (In GARP reading it is completely given ) and then next step of order the return set pairs is shown . ( ranked return of X).

Compute VaR and expected shortfall using the POT approach, given various parameter values.

87.1 If X is a random i.i.d loss with distribution function F(x), and (u) is a threshold value of X, what function defines the peaks-over-threshold (POT) approach?

a. F(x) = P{ X <= x | X > u}

b. F(x) = P{ X <= x | X = u}

c. F(x) = P{ X - u <= x | X > u}

d. F(x) = P{ X - u <= x | X = u}

87.2 Assume the following GP parameters under POT approach to extreme values: scale (beta) = 0.9, shape/tail index (xi) = 0.15, threshold (u) = 4.0%, and the percentage of observations above the threshold (Nu/n) = 10.0%. What are, respectively, the 99.5% and 99.9% value at risk (VaR)? (note: variation on Dowd's Example 7.5)

a. 3.95% (99.5%) and 5.24% (99.9%)

b. 4.15% and 6.24%

c. 7.40% and 9.97%

d. 9.03% and 11.31%

87.3 Using the same assumptions and same POT approach (generalized Pareto distribution), what are, respectively, the 99.5% and 99.9% expected shortfall (ES)?

a. 7.40% (99.5%) and 9.97% (99.9%)

b. 9.06% and 12.08%

c. 10.22% and 14.65%

d. 12.62% and 16.68%

87.1 C. F(x) = P{ X - u <= x | X > u}

Conditional on X exceeding the threshold (X>u), what is the probability that the loss in excess of the threshold (X-u) is less than or equal to x (i.e., CDF).

… note that F(x) is the parent distribution.

87.2 C. (7.404 @ 99.5% and 9.972 @ 99.9%)

See spreadsheet

87.3 B. (9.06 @ 99.5% and 12.08 @ 99.9%)

See spreadsheet

I might have discovered a typo in Tuckman chapter 8, study notes p. 42.

- Current calculation approach of one-year forward discount facor at node 1,1: 50%*[(1/1.1840) + (1/1.1040)]/1.10 = 0.766371
- How it should be at node 1,1: 50%*[(1/1.1840) + (1/1.1040)]/
**1.142**= 0.766371

- Current calculation approach of one-year forward discount facor at node 1,0: 50%*[(1/1.1040) + (1/1.0240)]/1.10 = 0.886233
- How it should be at node 1,0: 50%*[(1/1.1040) + (1/1.0240)]/
**1.062**= 0.886233

Thanks and best,

I tried to search the forum to see if this was previously discussed but couldn't find anything specific. Apologies if this is already discussed.

The learning spreadsheet for Dowd Ch3 and Ch4 does not seem to contain any of the examples used in the Study Notes/Learning Video. For instance, page 17 of the Study Note references return data for Apple, Applied Materials and HP Inc. for 21 days. I'm trying to follow the bootstrap method by replicating David's steps but I cannot find where this data is stored in the learning spreadsheet.

Thanks,

Evelyn

(in the meantime for Evelyn: working XLS to go here: tbd)

## Stay connected