Exam Feedback November 2016 Part 2 Exam Feedback

[Moey]

... also re gimme questions that unfortunately weren't asked - QQ plot, neyman-pearson / minimum risk and error / minimax, ordinal rankings such as spearman/tau - the lack of gimmes + long AND irrelevant item sets made this exam that much harder! Also don't think there were any plug n' play MVAR/CVAR questions like this forum mentioned last exam.

 

[FRMPART2]

One question was var given by historical simulation is lower or greater than by bootstrap historical simulation

 

[emilioalzamora1]

with regard to CoCo's: this is a term enshrined in the Basel III regulation since many years and everyone who wants to become a FRM should at least know what a CoCo is, irrespective of whether CoCo's are dealt with in great detail in the AIM.
I fully agree that the question around (CoCo's effect on capital requirements) was extremely harsh, but the term itself should be clear.

Why should questions be leaked beforehand? Who is telling this? Is there any proof?

 

[Moey]

with regard to CoCo's: this is a term enshrined in the Basel III regulation since many years and everyone who wants to become a FRM should at least know what a CoCo is, irrespective of whether CoCo's are dealt with in great detail in the AIM.
I fully agree that the question around (CoCo's effect on capital requirements) was extremely harsh, but the term itself should be clear.

Why should questions be leaked beforehand? Who is telling this? Is there any proof?

Emilio: COCO's are fair game especially w/ the uptick of their usage, but I also thought it was harsh to ask so many questions about them w/ little-to-no mention of them in practice exams or third-party questions. Re the possibility of cheating/leaking questions, I have no proof and didn't mean to raise alarm - just thought it was concerning what frmpart2dan said: "I am very surprised there are these questions discussed in Chinese forum but found nowhere else".

... also wanted to change gears from that mvar/ivar question.

 

[Nicole Seaman]

Here is a great post that David posted after the May exam for those of you who are interested in the historical pass rates: https://forum.bionicturtle.com/threads/may-2016-part-1-exam-feedback.9609/page-13#post-43490.

Nicole

 

[mghm]

CoCos are like Mezzanine tranche which can behave like equity when firm value falling and behaves like debt when firm value is rising. Difference in Mezz and Cocos - Mezz tranch is certificate against underlying and CoCos are instruments directly issued by financial institutions or Banks

 

[hsitaolee]

There are 5 item sets, and each has 3 questions. CCP was really heavily tested as one of the item set questions.
A lot of info in the item set was redundant. I skipped all of them and went back with Qs directly- and it turned
out my strategy worked well.

Hopefully the pass rate could be over 55% this time- the exam was really really tough!

 

[adambschrier]

does anyone remember the specifics on the CoCo questions/answers?

 

[equanimity]

One question was var given by historical simulation is lower or greater than by bootstrap historical simulation

I believe the answer to this question was that neither approach relies on a normal distribution. Did anyone else select that answer?

 

[frmpart2dan]

There are 5 item sets, and each has 3 questions. CCP was really heavily tested as one of the item set questions.
A lot of info in the item set was redundant. I skipped all of them and went back with Qs directly- and it turned
out my strategy worked well.

Hopefully the pass rate could be over 55% this time- the exam was really really tough!

If I remember correctly there were 6 item sets and at least one of them have 4 qns. I think this is the testing year when they start to put these item sets questions into the exam. For CFA there are a number of questions that are not counted (although you would not know which ones) and were put there for testing of later years exam, could it be some of the item sets are not included in the marking (perhaps the CoCo?!)

 

[Johnkrause1]

does anyone remember the specifics on the CoCo questions/answers?

I think one of the answers to one of the Coco questions is that it can be converted into equity when equity prices recover

 

[farahm]

The netting qns I picked 10, because I think it was asking the collaterals can be reduced. For example if A posts 7 collateral and B posts 5 collateral, then the total collaterals can be reduced is 5.

Correlation qns I went for receive more than 0.4 million because the firm is the buyer and the actual corr > forecasted and (forecasted - fixed) * notional = 0.4 million alr, so the buyer should receive more.

For the risk that is not mentioned I chose counterparty risk. Funding risk is related to obligation coverage which was mentioned I think, it said the investor worried that the fund cannot cover the liability.

Binary tree pricing is 945 usd smth.

Ho lee model is 5.1%

i remember being confused with binary tree model question - even though it was really easy. They asked for the current value i believe which made me confused between expected discount value or market value. which one was it?

 

[farahm]

I choose funding liquidity risk as liabilities are greater than assests and pensions funds will face problems to fund their pensions

I chose the same - I had that exact reasoning.

 

[farahm]

I hope everyone here passed, the fact that we know how to analyze the questions that much and agree on many things show how competent we are with the humongous syllabus of thd l2 FRM. Given the time and exposure to the knowledge, I think we all can make the right choice in real world

well said Dan.. just reading through all the comments, everyone seems to be defending their answers. on the other hand, quite scary how for a particualr question everyone had a different answer. very tricky exam !

 

[farahm]

I believe the answer to this question was that neither approach relies on a normal distribution. Did anyone else select that answer?

I forgot my answer , but I agree with the statement - both bootstrap and HS dont have to be rely on normal distribution.

 

[Trueman]

Why should questions be leaked beforehand?

I think a level of GARP practice exam should be appropriate.
If GARP presents to me for a training the question about capital adequacy like #80, where I just have to divide 145/3480, but at the actual exam I see series of questions about capital like these (CoCo) - I think it's not quite fair play.
I don't refuse to solve difficult questions, but I want to be ready for them

 

[Ra10as12]

There was one question related to expected shortfall at 99.5% CF ,does anyone remember that question ? we have to take avg from 99.6% to 99.9% to calculate ES at 99.5% Right ?

 

[Moey]

There was one question related to expected shortfall at 99.5% CF ,does anyone remember that question ? we have to take avg from 99.6% to 99.9% to calculate ES at 99.5% Right ?

Yes - one of the quick n' easy ones!

 

[fjc120]

Does anyone feel great about the exam? If yes, what and how did you study in terms of materials and practice exams? I read the entire books but they are extremely not straightforwad as BT helps tremendously but this time this exam had questions not even BT had covered as much. Hoping I have a ball bounce my way but i wont know best how to prepare. Does anyone recommend other materials that Can i can combine with BT? Is schweser worth a read? Wiley?

Dont worry i still think BT is best for it held my hand for part 1. Thanks david!

 

[raya]

There was one question related to expected shortfall at 99.5% CF ,does anyone remember that question ? we have to take avg from 99.6% to 99.9% to calculate ES at 99.5% Right ?

99.5% isn't taken into account? I thought we must compute the mean of var where var ≥ 99.5% to obtain ES