On L1 one of the questions was about box spreads.  Where in the AIMs is there anything about box spreads?  Its in my Hull book, Sixth Edition, Chapter 10, but it is not in the AIM.

Hi guys,

Here in Melbourne, we were 10 for L1 and about 10 for full exam. I did Level 1. Under normal exam conditions, I may fail with 99% confidence over 7 weeks time period. I believe I will definetely pass next time )) No worries. Without David’s fantastic materials, I wouldn’t even imagine the possibility of passing FRM exam. Good on you David, much appreciated. By the way, the best thing about FRM exam was the party I had last night. Cheers everyone. It’s all over.

It has been 24 hours since the exam and I was dreaming about solving quant questions just before I woke up!!!

Apart from the pass/fail considerations I found it to be a very intelligently structured exam. It made one think and some questions were actually amusing. I liked the clever use of Macguffins in the questions. The typos were of course an irritant!!!

Questions which stuck in my mind:

1. Where they asked you to rank the tracking errors by looking at graph of returns. I guess that is how most people do it in real life!!!

2. Where they asked you to estimate probability of right answers by guessing. Given the context, it was funny.

3. The HUGE amount of details they gave for a Historical Simulation problem where you just have to look the answer up!!!

4. Again the details given for the Maximum Likelihood estimation question. I was scared seeing it but read it fully anyways and found it was just testing if one knows the meaning of ML and ability to take log!!(which I did!!)

5. The innocous slipping in of the ‘lockout period’ in the credit card ABS paydown question which made me agonise no end (I have structured CMO’s for a living so it was a point of honour not to get any securitization questions wrong!!!!)Thankfully my understanding of lockout was correct(the correct answer was 0)

6. I had time only to do the L1 practice test from GARP site and it paid off somewhat as the LT volatilty question based on the GARCH equation was a straight lift from there!!!!!

7. I loved the large amount of practice oriented questions and questions dealing with the current crisis. The TED spread question really made me think as I was follwing the TED spread daily last November. It was very annoying that I still couldn’t answer that with any degree of confidence (I chose the option that said it didn’t rise immediately after Lehman but rose then on perception of higher CP risk)

I found the AM questions more interesting while PM ones more calculator oriented. I didn’t guess at all in the AM thus attempting only 49 but sanity prevailed in the PM test!    I wasn’t very well prepared so am expecting 60-65% as . Let’s see if that proves enough.

I do wish they allowed one to keep the question paper. It had plenty of food for thought. I doubt that any trainer could have thought up such a question paper!!!

I think there is no negative marking?.right?

Hi David, 

these are the questions that i still remembered (from L1) :

related to Storage Cost :
S0 = 320
r=0.5 % per month
storage cost (pay at the end) = 10 evey month

calculate the F0 for 1000 bushel in two months ?

related to GARCH (1,1) :
-ask about Long Run Volatility (not variance)
-ask about which has the long mean reversion (simply just ranking the alpha + Beta)

related to BSM
-given all parameter calculate r
-ask about Put ( i simply using Put-Call parity from the question before to get the ans)

There is no questions related : Michael Ong (UL & EL) & Operational Risk (Top Down vs Bottom Up)

after all, i don’t think i can pass the exam because too many guessing especially in Quantitative part (except Bayes & GARCH (1,1) where i found is easy for me).

Wrote full FRM exam in Mumbai. I found it good, nicely structured and balanced across study material. Its crafted to make sure that nobody passes by fluke or with little preparation.

I am sure to pass the exam with 95% confidence level with mean of 74% with standard deviation of 9%  ... Inspite of that my AIM to write FRM was never to look for certification but to set a traget for myself that would help me to read the concepts as some money would be at stake and hence will give me some push I wouldn’t have study half of this in 1 year that I did in 6 weeks without setting myself this target.

David, Thanks for the material and would like to contibute from here on….

TC

Can anybody body shed some light on these two :

related to Storage Cost :
S0 = 320
r=0.5 % per month
storage cost (pay at the end) = 10 evey month

calculate the F0 for 1000 bushel in two months ?

for this one I know either we devide storage cost to spot and get as a percentage or Present value way however I could not get any answer which was given??? :(

Regarding the BS put call parity? can we use it when we have one european call the the put is american?
thx

chiranjiv - 21 November 2009 08:12 PM

It has been 24 hours since the exam and I was dreaming about solving quant questions just before I woke up!!!

7. I loved the large amount of practice oriented questions and questions dealing with the current crisis. The TED spread question really made me think as I was follwing the TED spread daily last November. It was very annoying that I still couldn’t answer that with any degree of confidence (I chose the option that said it didn’t rise immediately after Lehman but rose then on perception of higher CP risk)

I found the AM questions more interesting while PM ones more calculator oriented. I didn’t guess at all in the AM thus attempting only 49 but sanity prevailed in the PM test!    I wasn’t very well prepared so am expecting 60-65% as . Let’s see if that proves enough.

I do wish they allowed one to keep the question paper. It had plenty of food for thought. I doubt that any trainer could have thought up such a question paper!!!

Hey…

I searched on Internet and found that TED spread rose immediately and retreated upon central banks’ intervention.

Also some questions I remember are

1. Conditional probability of economy remaining poor and neutral is 38.67% (ANS)
2. In global fund of funds hedge strategy, short selling is not uncorrelated with equity, it is -vely correlated (hence wrong option)
3. Short hedge benefits from strengthening of basis
4. Beta 1 and Beta 2 values (option 3 was what I was getting…If I remember correctly Beta 2 was 0.714)
5. The manager who wants to protect EU w.r.to US $ at 1.34 spot shouldn’t sell a call option as when it goes 1.34 sharply, incurs huge losses
6. I couldn’t get the LVAR answer but guessed that it should be above 1473 (VAR) but there two options. So I went for 1569
7. Home age is the only one that doesn’t cause pre payments

I don’t remember others….will post when I recollect

A few more bits of information about the L1 exam that I just remembered - that may or may not prove useful to others:

1. We were given a a table of probability values for the standard normal distribution (that needed to be used for some hypothesis testing).
2. One question required a Student’s T test - and they gave 4 values to choose from (you had to choose the correct degrees of freedom and confidence levels).
3. There were several different versions of the exam. The seats in front of me and to my left were both empty - and the barcode on the front page of those two exams had different endings (e.g. something like FRM-L1-3 and FRM-L1-4) whereas mine was “-2” on the end. I guess this is so that people can’t cheat by looking at their neighbours answers. I imagine though that they are the same questions but just in a diferent order.

I guessed most of the questions in AM session…was more comfortablw with the PM session….......

There was one question on Garp code of conduct….which i think most of you would agree was highly ambigous…..None of the options were looking right and all of them were looking wrong….hope you understand what i mean…Temporary suspension as well as permanent suspension both were part of answers…

David, can you give an idea of what could be the cutoff?( based on past papers, threshold…or comments about this paper given by everyone)...

Thanks..

Hello,

Full exam taken in Geneva.

I hope I pass, but “I should not sell the bear skin before killing as we say in french”.
It is all the most true as I had to guess several questions because I would have lost too much time on it.
The key risk parameter for this exam is TIME management.

As the questions are still fresh in my mind and in order, for future participants,
to gauge the exam difficulty,  here are the questions I remember.
I think that examples are the best way to know how tough are the questions.

Sorry for the spelling error, I am rushing as I was rushing during the exam…

1. Gold forward arbitrage with continuous yield (2%) and Rf (4%).
The future price was lower that the computed one.
What should be done to “gain” the arbitrage?

2. One BII market risk capital charge computation, be careful daily var was 95%.
So we had to convert to 99%.

3. Currency forward arbitrage (Rf and time provided)

4. Volatility smile descriptive question (currencies and equities)

5. Maximum likelihood (you should take the log and tell what to maximize/minimize)

6. No Poisson question

7. Easy Binomial: What is the prob that a student has less than 8 answers right
if he answers randomly a two possibility answer test with 20 question.
You did not have to compute but just to say how to compute.

8. Several questions that required put/call parity (including S*exp(-y*t))

9. Several questions on option combinations (butterfly spread)

10. One greek question

11. No TRS question

12. Several CDS questions (impact of correlations)

13. Lots of var questions

14. Not too many questions on BII compared to var questions.

15. One Tier1 and Tier2 BII computation

16. Model risk question

17. One UBS (easy) question

18. Why stress testing var

19. Indirect question on backtesting var

20. One easy swap computation (2 years remaining, Libor against 8%)

21. One easy question on interest rate: Which one is higher?
They gave yearly, monthly, quarterly and continuous numerical values.
The answer was continuous.

22. Several questions on linear regression:
Compute R^2. Beta has bad t-stat and the strategy is to be market neutral -> what can be concluded?

23. A few questions on IR, Sharpe and Treynor ratio.
I was confused by one because they were calling the tracking error, the systemic volatility or something like that.

24. One portfolio credit risk model question

25 One very easy question on matrix transition: Which statement is incorrect?
There was a line with BBB down grade proba > 80%...

26. No rating question

27. Difference between CDO and MBS (I answered the tranching)

28 CDS question on price of first to default versus second to default.

29. No moral hazard question

30. One SPE usage question.

40. No inverse floating question

41. A few modified duration question. (DeltaV = -D* * V * deltaY)

42. Which obligation has negative convexity:
good to see easy question and not time consuming.

43. An easy proba computation:
z-proba of not been between 1 and 1.5 or something like that.
The z table was provided for each test on the first page.

44. P(A|B) computation:
I had P(A and B) by using P(B|A) but I did not have the time to compute P(B)
because the text was long and I needed at least 5 minutes to recompute.
This is a typical question when I lost point because I had to avoid loosing time.

45. One easy question kurtosis:
The normal has lowest proba of extreme value than 4, 8 kurtosis distributions.

46. Several EVT questions: one was to compute it with extreme returns provided.

47. One var question:
20 worst returns provided but unsorted. Find the var, It was 1% of 80 Million as far as I remember.

48. You decrease significance what happens to Type I and Type II probs.

49. One binomial tree call computation question that I failed.

50. One Monte Carlo simulation of GBM (mu was zero to be easier), sigma was provided
You had to compute S(n+2) given S(n) e(n) and e(n+1). 
I also failed because I did not want to loose too much time on computations
and I made a mistake in the contribution.

51. No Cholesky or interest rate model question.

52. One Ted behaviour (after Lehman episode) question.
But there were two possible good answers as far as I understood.
I answered that the it increases because there was a lack of liquidity
(for me the fact that all bankers we scared of lending was a liquidity funding problem).

53. One or two linear hedging question (rho * sigma(S) / sigma(F))

54. Concerning study cases, a single question on MetallGesellshaft

55. Two questions on basis risk.

56. A liquidity VAR computation:
Normal distribution for return with mu and sigma, value of portfolio and spread value in dollar provided

57. One EWMA and one Garch (compute the long term vol for Garch)
This has nothing to do with the exam. But it remembers my wife’s boss explaining to me that Garch
has nothing to do with modeling heteroskedostaticity.
As he is responsible for Market Risk Management, I should have suggested him to take the Garp exam.

58. An interesting question on style drift versus leverage increase for HF.
They were providing the fund returns and the benchmark returns.
You could see that one manager increases leverage (same signs for returns) and the other was doing style drift.

58. Tracking error:
3 graphs were given with benchmark and fund returns. Tell which one has lowest tracking error.

59. Correlation:
3 graphs were provided. Which group of variables has the highest correlation?

60. An easy question on credit risk: Which operation adds CR: sell/buy options

61. One Merton question (that I failed) close to the one provided in Garp2009:
Compute prob of default

61. Two EL and UL computations (PD, LGD, COM, sigma(PD), sigma(LGD) were provided)

62. Two questions on credit “modifiers” (Triggers, Netting, collateral, MTM)

63. No cat bound question

64. One Raroc computations (no taxes were involved and RC was provided)

65. One BS call/put computation N(d1) and N(d2) were provided

66. One interest risk reduction question (buy a cap, sell a floor, ...)

67. One or two questions on MVAR, CVAR questions:
Compute the global VAR if a fund is removed
Compute an MVAR

See you,

Good luck everyone!

Hi all,

The AM paper took up more time… but the PM one gave less of a headache. I heard people and even a colleague complaining about the difficulty of the paper. Quite a lot of people missed the exam so it is based on 50% of those who turned up?

Anyway thanks David for your great help! You’ve been so great! But I’m really scared off… I don’t think I’l try again if I fail. Put in nearly 800 hours of study and still no tsure whether can pass. :( Maybe 55% confident.

benoit16 - 22 November 2009 05:59 AM

Hello,

Full exam taken in Geneva.

I hope I pass, but “I should not sell the bear skin before killing as we say in french”.
It is all the most true as I had to guess several questions because I would have lost too much time on it.
The key risk parameter for this exam is TIME management.

As the questions are still fresh in my mind and in order, for future participants,
to gauge the exam difficulty,  here are the questions I remember.
I think that examples are the best way to know how tough are the questions.

Sorry for the spelling error, I am rushing as I was rushing during the exam…

1. Gold forward arbitrage with continuous yield (2%) and Rf (4%).
The future price was lower that the computed one.
What should be done to “gain” the arbitrage?

2. One BII market risk capital charge computation, be careful daily var was 95%.
So we had to convert to 99%.

I think this was coming to around 267000 or close to that after scaling the 95% Var to 99% and to 20 days

3. Currency forward arbitrage (Rf and time provided)

4. Volatility smile descriptive question (currencies and equities)

5. Maximum likelihood (you should take the log and tell what to maximize/minimize)

I think here Lambda should be maximum. Second choice goes

6. No Poisson question

7. Easy Binomial: What is the prob that a student has less than 8 answers right
if he answers randomly a two possibility answer test with 20 question.
You did not have to compute but just to say how to compute.

This was option 3 as the second option was sounding right at the first sight but it was a uniform distribution

8. Several questions that required put/call parity (including S*exp(-y*t))

There was one question on European PUT and American CALL…...I htink we can’t apply put/call parity. So I said the same.

9. Several questions on option combinations (butterfly spread)

There was this box spread arbitrage. Is it option 3?

10. One greek question

I think in the money options lose more than any other option.

11. No TRS question

12. Several CDS questions (impact of correlations)

One of the CDS answers, I was getting +14 to the banker

13. Lots of var questions

14. Not too many questions on BII compared to var questions.

15. One Tier1 and Tier2 BII computation

This was option C in my paper. Don’t know if they jumble the answers as well.

16. Model risk question

17. One UBS (easy) question

I think they were easily earning over LIBOR by investing AAA.

18. Why stress testing var

19. Indirect question on backtesting var

20. One easy swap computation (2 years remaining, Libor against 8%)

I got it as +14 to the bank.

21. One easy question on interest rate: Which one is higher?
They gave yearly, monthly, quarterly and continuous numerical values.
The answer was continuous.

Yeah. Even I got the same answer.

22. Several questions on linear regression:
Compute R^2. Beta has bad t-stat and the strategy is to be market neutral -> what can be concluded?

One was coefficient of determination. Given Beta was .977. So the R^2 turns out to be .913

23. A few questions on IR, Sharpe and Treynor ratio.
I was confused by one because they were calling the tracking error, the systemic volatility or something like that.

24. One portfolio credit risk model question

25 One very easy question on matrix transition: Which statement is incorrect?
There was a line with BBB down grade proba > 80%...

26. No rating question

27. Difference between CDO and MBS (I answered the tranching)

Don’t we have tranching in both? I think we do….but not sure.

28 CDS question on price of first to default versus second to default.

This, the FTD always is costlier or equally costly to STD second to default

29. No moral hazard question

30. One SPE usage question.

40. No inverse floating question

41. A few modified duration question. (DeltaV = -D* * V * deltaY)

There wasn’t a perfect answer .....I went for the closest one.

42. Which obligation has negative convexity:
good to see easy question and not time consuming.

43. An easy proba computation:
z-proba of not been between 1 and 1.5 or something like that.
The z table was provided for each test on the first page.

44. P(A|B) computation:
I had P(A and B) by using P(B|A) but I did not have the time to compute P(B)
because the text was long and I needed at least 5 minutes to recompute.
This is a typical question when I lost point because I had to avoid loosing time.

This was the one I was talking about. It came out to be 38.67%

45. One easy question kurtosis:
The normal has lowest proba of extreme value than 4, 8 kurtosis distributions.

Yeah. Even I went for it.

46. Several EVT questions: one was to compute it with extreme returns provided.

47. One var question:
20 worst returns provided but unsorted. Find the var, It was 1% of 80 Million as far as I remember.

Exactly. I got the same.

48. You decrease significance what happens to Type I and Type II probs.

TYPE II increases

49. One binomial tree call computation question that I failed.

I wasn’t getting answer for this. I got less than 1. So I went for the lowest answer which was 2.

50. One Monte Carlo simulation of GBM (mu was zero to be easier), sigma was provided
You had to compute S(n+2) given S(n) e(n) and e(n+1). 
I also failed because I did not want to loose too much time on computations
and I made a mistake in the contribution.

There were two options like 97.79 and 99.79. We also had 99.69. So I thought it should be 99+ and should .79 => 99.79

51. No Cholesky or interest rate model question.

52. One Ted behaviour (after Lehman episode) question.
But there were two possible good answers as far as I understood.
I answered that the it increases because there was a lack of liquidity
(for me the fact that all bankers we scared of lending was a liquidity funding problem).

This sounds very much logical but one of our collegues on this site has posted a link which says it peaked in Oct.

53. One or two linear hedging question (rho * sigma(S) / sigma(F))

54. Concerning study cases, a single question on MetallGesellshaft

THis was backwardation to Contango and not the otherway round

55. Two questions on basis risk.

Short hedge benefits from basis strengthening. The other one I think only that had same underlying and same T will not have Basis risk and all others had.

56. A liquidity VAR computation:
Normal distribution for return with mu and sigma, value of portfolio and spread value in dollar provided

I wasn’t getting answer for this.

57. One EWMA and one Garch (compute the long term vol for Garch)
This has nothing to do with the exam. But it remembers my wife’s boss explaining to me that Garch
has nothing to do with modeling heteroskedostaticity.
As he is responsible for Market Risk Management, I should have suggested him to take the Garp exam.

THere was another. THat has increasing R^2. Is it coz of multicollinearity?

58. An interesting question on style drift versus leverage increase for HF.
They were providing the fund returns and the benchmark returns.
You could see that one manager increases leverage (same signs for returns) and the other was doing style drift.

This was a good logical question. I answered the same way.

58. Tracking error:
3 graphs were given with benchmark and fund returns. Tell which one has lowest tracking error.

The one which had two graphs move parallelly was what I have chosen. It may not be correct.

59. Correlation:
3 graphs were provided. Which group of variables has the highest correlation?

60. An easy question on credit risk: Which operation adds CR: sell/buy options

selling a call option I guess

61. One Merton question (that I failed) close to the one provided in Garp2009:
Compute prob of default

I got DD as 3.6

61. Two EL and UL computations (PD, LGD, COM, sigma(PD), sigma(LGD) were provided)

62. Two questions on credit “modifiers” (Triggers, Netting, collateral, MTM)

63. No cat bound question

64. One Raroc computations (no taxes were involved and RC was provided)

I was getting it as 2.6 to achieve a min of 15% return

65. One BS call/put computation N(d1) and N(d2) were provided

66. One interest risk reduction question (buy a cap, sell a floor, ...)

I went for options i, ii and iii.

67. One or two questions on MVAR, CVAR questions:
Compute the global VAR if a fund is removed
Compute an MVAR

See you,

ajsa - 22 November 2009 06:07 AM

about the TED spread:
http://www.crystalbull.com/stock-market-timing/TED-Spread-chart
LEH filed bankruptcy on 9/15, while TED spread peaked on late Oct

I see that it indeed peaked immediately. I think we should check for 2008 and the dates are so close on the graph.