Exam Feedback May 2018 Part 1 Exam Feedback

[Stuart D Moncrieff]

There was a question that seemed to be as follows (the numbers are wrong but the idea is the same, I think):

There are three grades of bonds. AA, BB and C. P(default) for each of them is 0.1, 0.25 and 0.4. If a bond defaults, what is the probability that it is a BB or C bond?

The way I calculated it was as such: (0.25+0.4)/(0.1+0.25+0.4).

Please correct me if I am wrong. Also please correct me if I am remembering the question wrongly.

------------------------

Another question on margin requirement. I got the loss on Day 1 as 54,000. As the initial margin is 8000*4 = 32,000, this meant the balance at the end of Day 1 was 32,000 - 54,000 = -22,000.

What is the answer for the status of the account on Day 2? I put it as a variation margin of 54,000...is that correct?

-----------------------

There was a question on hedging with DV01. I can't recall which I used as the denominator, but I got a hedge ratio of 0.5, therefore half the amount of the current position required to hedge. In my case, I believe the current position was 48mil and so the hedge amount was 24mil (could be wrong with the figures but this is the idea).

-----------------------

There was a question to find the std dev and covariance of returns of three data that were provided. I used the formula as SUM[(X - Xbar)*(Y - Ybar)]/(N-1) for the covariance and a similar formula for the std dev. I know the answer I got was 0.04 for std dev and 0.0016 for covariance. However, the other answer was 0.03 for std dev and 0.11 for covariance. For this question, was the denominator for the std dev and covariance just supposed to be N or N-1? In my case I took it as N-1 but am not certain about it.

I thought that since the price remained unchanged on the second day the variation margin should have been 0 since the margin should be updated daily. I wasn't certain though.

For the probability question I remember I constructed a table like David used most in his examples - method was similar to you but I think I might have applied the odds of it being an AA, BB or C bond on top of the probability of defaults. Can't quite remember the wording of the question but I remember constructing the table.

 

[nikic]

yes, PD*EAD* LGD (1-recovery value).... that´s expected value

I remember one of rerunning a previous sample or something like that, answers were antithetic, control variates or bootstrapping. I put bootstrapping as the correct one. ¿any feedback?

I put bootstrapping too but I checked the book after the exam and it was antithetic variate if I am not mistaken. Sadly this was a stab in the dark for me.

 

[pranay1986]

I put bootstrapping too but I checked the book after the exam and it was antithetic variate if I am not mistaken. Sadly this was a stab in the dark for me.

Antithetic referred to the complement set .. and i think the question referred to the complement set correct?

 

[nikic]

I thought that since the price remained unchanged on the second day the variation margin should have been 0 since the margin should be updated daily. I wasn't certain though.

For the probability question I remember I constructed a table like David used most in his examples - method was similar to you but I think I might have applied the odds of it being an AA, BB or C bond on top of the probability of defaults. Can't quite remember the wording of the question but I remember constructing the table.

Did the question give the odds of it being a AA, BB or C bond? I don't recall it. If it did give the odds, then I'm definitely wrong and wonder why I didn't go down that route. Unless I did and am not remembering it.

If the odds were given (say AA=50%, BB=30%, CC=20%)...is such the formula?

0.3*P(DefaultB) + 0.2*P(DefaultC) / [0.5*P(DefaultA) + 0.3*P(DefaultB) + 0.2*P(DefaultC)]

Now that I think about it I definitely used the proportions if it were provided.

 

[hana1563]

yes, PD*EAD* LGD (1-recovery value).... that´s expected value

I remember one of rerunning a previous sample or something like that, answers were antithetic, control variates or bootstrapping. I put bootstrapping as the correct one. ¿any feedback?

antithetic was the right answer

 

[jbejarjimenez]

There was one question on operational risk UL calculation EA was 80 mil recovery rate was 67% and probability of Default was 12%. i think none of the option was correct, did anyone else faced same issue ??

I had exactly the same problem, I think that I performed all the calculations right but I still didn’t find any answer that matched my numbers... really weird

 

[Stuart D Moncrieff]

Did the question give the odds of it being a AA, BB or C bond? I don't recall it. If it did give the odds, then I'm definitely wrong and wonder why I didn't go down that route. Unless I did and am not remembering it.

If the odds were given (say AA=50%, BB=30%, CC=20%)...is such the formula?

0.3*P(DefaultB) + 0.2*P(DefaultC) / [0.5*P(DefaultA) + 0.3*P(DefaultB) + 0.2*P(DefaultC)]

Now that I think about it I definitely used the proportions if it were provided.

Yeh I agree with you, I'm a bit unsure myself, I remember making the 2x2 table but I don't know why I would have done that if there weren't the numbers available to construct it....

 

[Albertikos]

And for the copula... what was the correct answer?

 

[nikic]

Finally Tuesday.... Let me download my memory

4. Futures price involving Cost of carry relating to storage cost per month paid in advance.
6. Some positions by various Counterparties to work out which one has the greatest exposure.
8. Optimisation of Shareholders wealth.
11. How banks plan for EL in their products.
19. Spot rates to calculate Forward rate with annual compounding. Almost same as practice question.
20. GARP code of conduct question on use of previous employers data for current employer due to time.
25. Calculate the outstanding mortgage for a particular year remaining from a set of given values e.g mortgage value rate etc
26. SMM or CPR calculation with initial value at begining of month and what value was paid off.
28. Think a question around Risk Management failure by not using correct distribution.
29. ERM question was very tricky. I think it asked about Line Management/ Risk Transfer role on the Stack or responsibilities.
Can't remember seeing anyone on Key rate shift. Time wasnt enough as usual coupled with the delayed start at Ibis London.

Hey Dotun, my queries on the questions highlighted by your good self as follows:

4) How did you answer this question? Is it as simple as S*(e^rt) + Cost + Cost*(e^-rt/12)? I believe the storage cost was $6. My answer was 1242 or something. Basically around $12 above the vanilla futures price. The answers were spaced $2 apart.

6) Would this be simply finding the counterparty with the highest negative exposure?

8) I don't recall this question. Can someone jog my memory?

11) I can't recall the answers. What's the correct answer for this?

19) I don't recall such a question at all. I do recall a given rate for semi-annual compounding and you were asked to calculate the corresponding rate for monthly compounding. Is this what you're referring to? There was another question with 1.5 year discount rate given and you were asked to calculate the corresponding spot rate.

20) Is the answer something along the lines of "execute all services with diligence and perform all work in a manner that is independent from interested parties"? The others sounded wrong.

25) Sadly don't recall this question!

26) Is this the one on prepayment? Find the monthly repayment, and minus that from the amount that was paid for the particular month to find the pre-payment amount?

28) Not able to recall this either!

29) Is this the one where the answer contained the words "line management" as well?

Thanks!

 

[jshi]

yes, PD*EAD* LGD (1-recovery value).... that´s expected value

I remember one of rerunning a previous sample or something like that, answers were antithetic, control variates or bootstrapping. I put bootstrapping as the correct one. ¿any feedback?

I think it was antithetic as I recall the question specifically called out "complementary" set of inputs or something of that sort.

 

[nikic]

And for the copula... what was the correct answer?

This is the one on V1 and V2 correct? I believe I answered Gaussian copula but it was a total guess. Looking into the reference book, it seems to be correct?

 

[Albertikos]

This is the one on V1 and V2 correct? I believe I answered Gaussian copula but it was a total guess. Looking into the reference book, it seems to be correct?

I put the same...

 

[jshi]

And for the copula... what was the correct answer?

Ah.. I can't remember exactly what the question was, something about 2 normal distributions so I chose C/the one on Gaussian copula. Not very sure... I found the paper having so many difficult qualitative questions on items that's not a big LO

 

[jbejarjimenez]

There was a question asking about what role played Duraton and convexity in case of increase in rates. A marked a decrease in price due to duration and an increase due to convexity... what did u answer? Thnks

 

[nikic]

There was a question asking about what role played Duraton and convexity in case of increase in rates. A marked a decrease in price due to duration and an increase due to convexity... what did u answer? Thnks

This is correct. Convexity is always positive due to the square, duration has the opposite effect.

On that note, there was a question to calculate portfolio duration and convexity. I got the duration easily...but the convexity alluded me. What did I miss out on? I believe there were two answers - 14.3 and 16.4. At least one was 14.x and another 16.x. I took a gamble on 16.x. No idea!!!

 

[nikic]

Some questions i remember about information criterion with best consistency answer was SIC

expection of jump on both sides for a security which is the best spread to apply to capture volatility using puts
I marked long butterfly answer is i think is short butterfly

One question on difference between future and OTC contract

Operationl risk allocation

One on korean bank recieving AUD dollars best hedge to prevent some risk but also gain if its in favour i marked long put

Anyone remember the nature of the question on the futures/OTC and the correct answer?

Also for the operational risk allocation, is this the one where there were two answers talking about the basic indicator approach?

For the currency hedge using options, I believe I put long put as well (it gains if the currency goes the wrong way). Was this correct?

 

[nikic]

For the one of the information ... the manager says to select the one with the highest information ratio but at least 0,30 sharpe.

I remember another one which was literally one sheet of the exam of a multi factor question.

Yes, the multifactor question. Is it a case of simply plugging in the numbers into the respective formulas? I got 6.5% and 6.65% respectively or something very similar to that (Answer A). Or is there something else that needs be done?

 

[jshi]

What's the difficulty of this paper compared to previous sittings?

What would an estimated passing mark be like? Would 60++ be enough?

I have not taken any other sittings. I read the other exam threads and it seemed to be similar sentiment that the exam is always more difficult than the GARP exams or the providers. I personally found it a very difficult challenge in the time pressure and for some questions, more time would not even been able to help me as I was almost completely stuck on some. I was told in most cases to aim for 70%++ to pass, I feel like I am below that mark I had about 20 questions I had to skip and go back to to, out of which maybe only 10 I felt okay about. The remaining 80 was a mix of good guesses, confident matching answers and ones I knew I wouldn't get the answer anyway and made a random choice. I usually score about 65-70% on Schweser and BT papers so this was a bit disappointing for me.

Quantitative questions it was difficult to guess as the numbers were close together or directionally the same, meaning you could only eliminate 1-2 options max. Also many irrelevant numbers and unsure what the question was asking. I can safely say almost all were multistep (size/value of contracts, FX, prior cost of borrowing or difference between options 1 and 2, variance of PD had to be calculating, EWMA and GARCH other questions required portfolio rather than single asset) and very few basic formula plug and play type (even the cost of carry wasn't the usual is there arbitrage you should buy the spot etc but rather on lease rates, convenience yields, the specific market, what this meant). I felt topic overlap and overwhelmed by the numerous hedging questions from T3 (hedge ratio and no of futures contracts) and T4 (Duration based hedging, DV01) that made me slow down so much and I ended up just skipping. My confident topics didn't necessarily become confident answers but at least gave me direction on where to continue my efforts after skipping the lengthy questions

Qualitative questions were also tricky and down to 1-2 specific words and also tested on the more obscure LOs (data producing methods of MCS, copulas) so reading the question made you go "Damn! This bit!"

 

[jshi]

Yes, the multifactor question. Is it a case of simply plugging in the numbers into the respective formulas? I got 6.5% and 6.65% respectively or something very similar to that (Answer A). Or is there something else that needs be done?

I think I had the same answer for this as well. I believe there were 2 that had the same numbers, and then the difference between the 2 statements was suggesting whether multifactor could predict the expected return / the market movement better or something.

 

[Albertikos]

Anyone remember the nature of the question on the futures/OTC and the correct answer?

Also for the operational risk allocation, is this the one where there were two answers talking about the basic indicator approach?

For the currency hedge using options, I believe I put long put as well (it gains if the currency goes the wrong way). Was this correct?

For 1 not sure, the was something related to the settlement or something like that.
For 2, put the one different betas for each business lune.
For 3, don´t remember what i put...

One that you have to calculate spot in the future... spot*exp((dc-fc)*t)--... was the answer less than the actual spot?