Course Errors Found in 2021/2022 Study Materials P2.T5. Market Risk

[Nicole Seaman]

Please use this new thread to let David and I know about any errors, missing/broken links, etc. that you find in the 2021/2022 materials that are published in the study planner under P1.T4. Valuation & Risk Models. This will keep our forum much more organized. We appreciate your cooperation!

PLEASE NOTE: Our Practice Question sets already have links to their specific forum threads where you can post about any errors that you find. This thread is for any other materials (notes, spreadsheets, videos, etc.) where you might find errors.

Information needed for us to correct errors:

• Reading
• Page number
• Error

 

[FRMNinjaLeonardo]

On p.59 of the notes for T5 R5 Tuckman's ch9, the model highlighted in blue is wrong

 

[blackbread]

Page 70 of the notes for P2.T5 Tuckman chapter 10, starting node of the interest rate tree diagram shows the natural logarithm of the rate rather than the initial rate itself.

 

[Ellie_10]

On page 34 of the notes for T5.R1 Dowd Chapter 7:

Describe the peaks-over-threshold (POT) approach

Formula 7.18:

I believe it should be Pr{X-u<= x | X > u}

 

[kchristo]

In T5-R1-P2-Dowd, under "Define coherent risk measures.", the definition of positive homogeniety refers to X and Y in the image but only X in the text. (P10)

 

[Branislav]

Dear @Nicole Seaman , regarding Hull, Chapter 18. Fundamental Review of the Trading Book study note, I think square root of ten is missing in VaR formula.

kind regards

 

[poojanmehta1]

@Nicole Seaman - Am looking at pdf notes of Bruce Tuckman, Fixed Income Securities. The blank page is under Tuckman, chapter 7. (page 22 of the pdf)

Also in the same chapter - Define risk-neutral pricing and explain its use in option pricing, in para one it mentions the example of coin toss for risk neutral pricing. Am assuming the payment on coin toss should be either $2 or $0 (mentioned as 1$)
"Let’s assume a fair coin (ie, 50% probably of flipping heads) while the risk-free rate is 3.0% with continuous compounding. You are offered the chance to buy a digital (aka, binary) option that pays, in one year when the coin is flipped, either $2.00 or $1.00"

 

[bollengc]

in Tuckman chapter8,
page 38 of pdf T5-R5-P2-Tuckman-(MR10-MR14)-v3
the power 3 should be outside of the brackets (1+r(3))^3

 

[Torsleno]

Hello,

Little typo in

Study Notes: Schroeck, Chapter 5: Capital Structure in Banks​

Define and calculate unexpected loss (UL).
Page 7



the sigma^2 should be variance of LR/PD (not standard deviation)

 

[cnomGreen]

Market risk practice questions when printed as PDF: The answer to 702.3 states that the answer is "D", although the associated value is answer "C"

 

[novenagates]

Hello, dear @David Harper CFA FRM @Nicole Seaman

For P2. T5 Tuckman's Study Notes, on Page 41

The denominator in both calculations seems wrong, we shouldn't use 1.10 in either case, instead we should use for

node [1,1], 1.142, the 14.2% is the right discount rate for end of Year 1
node [1,0], 1.062, the 6.2% is the right discount rate for end of Year 1

Similar error occur on the Slides for the Video tutorials for this chapter too, on Page 16

The calculations on the Excel spreadsheet are correct.

Thank you.

 

[enjofaes]

Page 8 market risk does not seem right for ES



If 97% ES = conditional average of the worst three losses
Then 95% ES = conditional average of worst five losses (like in formula not the text) OR 4 losses ?
96% ES = conditional average of worst 4 losses then -- again not like in the text. OR 3 losses?

 

[dla00]

Hi @David Harper CFA FRM , @Nicole Seaman

On part "Define coherent risk measures (continued)" from Instructional Video: Dowd, Chapter 3: Estimating Market Risk Measures.

a. Regarding positive homogeneity, I think it should read lambda * rho(X) I think, rather than Y.
b. Regarding translation invariance, I think the right hand side should be rho(X) - c instead of +c.

Regards

 

[David Harper CFA FRM]

Hi @enjofaes Yes, agreed, my text was a bit whacked. Thank you! I have edited it to reflect text that matches the exhibit, as follows:

If we seek the 97.0% ES, then we want the (conditional) average of the worst three losses: (64.28 + 63.27 + 36.41)/3 = 54.65 is the 97.0% ES because it is the (equally-weighted, in this case) average of the worst 3.0% of losses. If we seek the 95.0% ES, then we want the (conditional) average of the worst five losses: (64.28 + 63.27 + 36.41 + 36.27 + 35.89))/5 = 47.22 is the 95.0% ES because it is the (equally-weighted) average of the worst 5.0% of losses. Bear in mind this is a discrete distribution, but the same concept applies to a continuous distribution.

Put simply,

• 97.0% ES is average of 3 worst (3%) losses: (64.28 + 63.27 + 36.41)/3 = 54.65
• 96.0% ES is average of 4 worst (4%) losses: not discussed
• 95.0% ES is average of 5 worst (5%) losses: (64.28 + 63.27 + 36.41 + 36.27 + 35.89))/5 = 47.22

 

[David Harper CFA FRM]

Hi @dla00 Yes that's our mistake in the video, in both cases. It should be as reflected in this video https://forum.bionicturtle.com/thre...k-measures-and-why-var-is-not-coherent.22464/

 

[dla00]

Hi @enjofaes Yes, agreed, my text was a bit whacked. Thank you! I have edited it to reflect text that matches the exhibit, as follows:

Put simply,

• 97.0% ES is average of 3 worst (3%) losses: (64.28 + 63.27 + 36.41)/3 = 54.65
• 96.0% ES is average of 4 worst (4%) losses: not discussed
• 95.0% ES is average of 5 worst (5%) losses: (64.28 + 63.27 + 36.41 + 36.27 + 35.89))/5 = 47.22

There is also an extra parenthesis after 35.89.

 

[dla00]

Hi @dla00 Yes that's our mistake in the video, in both cases. It should be as reflected in this video https://forum.bionicturtle.com/thre...k-measures-and-why-var-is-not-coherent.22464/

Thanks @David Harper CFA FRM; just in case you want to update the notes: similar errors appear in VitalSource notes on p.10 [(iv) translation refers to "n" when it should be c and homogeneity refers to r.v Y when it should be X].

 

[David Harper CFA FRM]

Hi @dla00 Great, thank you for those additional comments re translation invariance and homogeneity. I was actually aware of those mistakes because, as part of updating the notes, we go through the cited errors and they have been surfaced. Thanks again!

 

[dla00]

Hi @David Harper CFA FRM

1) On p.135 of notes "Jorion, Chapter 6: Backtesting VaR", the second to last item says (1-p) > 10, it should read (1-p)*T > 10, so sample size is missing.
2) On p.136 of the same set of notes, "Over 250 days, a good 95% VaR model will produce approximately 5% * 500 days = 25 exceptions". Did you mean 500 days instead of 250?
3) On p.152 of the same set of notes, the exposures (in Table: Mapping Exposures) of the total PF are x_1, x_1 and x_1. In order to be consistent with the columns, did you mean x_1, x_2 and x_3?
4) On p.153 it says "Ignoring \alpha" but it should be "Ignoring \alpha_{i}".
5) Question: On minute 10:10 of Instructional Video: Risk Measurement for the Trading Book you mention it is in "P/L format". I find this confusing:
losses are to the right so if you take look at the tail (RHS) the corresponding quantile should be the VaR, so shouldn't be this L(+)/P(-) format?
Remark a similar distribution plot appears in official GARP book (Market Risk Measurement and Management) figure 1.2 on p.4 and the legend is displayed as L(+)/P(-) format.

Thanks

 

[David Harper CFA FRM]

Hi @David Harper CFA FRM

1) On p.135 of notes "Jorion, Chapter 6: Backtesting VaR", the second to last item says (1-p) > 10, it should read (1-p)*T > 10, so sample size is missing.
2) On p.136 of the same set of notes, "Over 250 days, a good 95% VaR model will produce approximately 5% * 500 days = 25 exceptions". Did you mean 500 days instead of 250?
3) On p.152 of the same set of notes, the exposures (in Table: Mapping Exposures) of the total PF are x_1, x_1 and x_1. In order to be consistent with the columns, did you mean x_1, x_2 and x_3?
4) On p.153 it says "Ignoring \alpha" but it should be "Ignoring \alpha_{i}".
5) Question: On minute 10:10 of Instructional Video: Risk Measurement for the Trading Book you mention it is in "P/L format". I find this confusing:
losses are to the right so if you take look at the tail (RHS) the corresponding quantile should be the VaR, so shouldn't be this L(+)/P(-) format?
Remark a similar distribution plot appears in official GARP book (Market Risk Measurement and Management) figure 1.2 on p.4 and the legend is displayed as L(+)/P(-) format.

Thanks

HI @dla00 Thank you! In regard to the four cited errors, I agree with you: all four are mistakes and I just fixed them. Thank you! That's super great feedback.

Re risk measurement video, I will take a look but I agree with GARP's Figure 1.2 (shown below): if losses are positive then the format is L(+)/P(-), and vice versa. It's entirely possible I misspoke. Thank you!