Exam Feedback May 2019 Part 1 Exam Feedback

[Nicole Seaman]

We hope that everyone did well on the FRM Part 1 exam on Saturday! We would love to hear any feedback that you have about the exam. How did it go? Did you encounter unexpected questions? Thank you in advance for any feedback you can provide!

Also, if you would recommend our study program to others, we would very much appreciate a review on our Facebook page!

 

[Rohanm94]

Found it tricky and lengthy. The theory questions were so close! Expecting 55-60/100

 

[Detective]

I felt it was very fair. It was easier than I expected it to be (far easier than the practice GARP exam in my opinion). There were computations, but far less involved. To give an example, the practice exam had a question on 2 step binomial tree, but on actual exam it was a 1 step question. Felt 4 hours was ample given questions that were being asked.

In my opinion, if some carefully read source readings and/or BT notes, and truly did every practice question BT has to offer and understood solutions, a 90+/100 is very tenable.

I didn’t get to do much practice as I’d hoped (basically went over all of Topic 4 - Valuation methods in ~2 days). Only did GARP practice exam and two mock BT exams (mock 1 & 2). Whoever put mock 2 together did a very good job. I felt there were 10 questions from that mock almost verbatim on the actual exam.

The biggest challenge to me was multiplication and long division as my calculator stopped working a little less than 10 questions into the exam. Even though the exam felt more qualitative, there were a decent amount of questions that required calculator. I could usually get it down to 2 choices from intuition, but lack of computing power did make me disadvantaged on around 10 questions (that involved sqrt or exp something and choices were too close together).

Some types of questions I remember:
-CAPM updated betas/factors and finding change in estimated return (David had lot of these questions)

-Updated covariance from EWMA. Very tedious problem even with a calculator. Surprised they tested it, but BT had a lot of these problems. I ended up just finding covariance today using yesterday’s correlation and deviations, and eliminating that answer (which was a possible choice). I can’t remember why, but I reasoned based on numbers today cov update would be higher and there was only 1 choice higher than wrong answer. However, I think I may have been wrong since they had 2 choices lower than the wrong answer.

-They actually asked on Sortino’s ratio. I didn’t bother memorizing the formula so wasn’t too bothered about lack of calculator.

-A good amount plug and chug lazy questions from GARP, with little to no insight required:
—Delta of put given N(d1)
—Unexpected Loss computation (try doing this one w/o a calculator...)
—Poison Process
—Binomial Probability came up on ~2-3 questions.

Misc:
—Sic is consistent and penalizes more heavily.
—ES is subadditive while VAR is not (this one I felt was worth reading the source since they had a 3-4 page build up to this point, so I remembered it. It was probably in BT notes as well, but I am less likely to remember).
—Interest rate parity came up. I didn’t remember formula, thought I’d be able to reason my way out of it, but only managed to get it down to 2 choices. With formula it should have been easy.

Still keeping out some hope that I passed, but I think it’s a toss up at this point.

 

[VWJETTY]

I agree with detective for the most part.

Although I somehow got stuck on the poission calculation towards the end even though I knew the formula for it. And ended up guess it (educated guess I would like to think). But other than that I'm just gonna give a brain dump here.

ES given the historical returns and 99% confidence internval. I think N was 200. So I just took the average of the lowest and the second lowest return? Is that right?

What was the difference between FRA and interest rate swaps? I didn't really review FRA at all so I was in the water on that one.

A good amount of stress testing, governance, quanlitative stuff. Geezzz If I could only remember them now

Oh yes, there was one about a teacher, and teaching the interms about proper governance, I said that the speculative bet would expose them to foreign currency risk. That was actually my last question.

Agree on the SIC being the heaviest penalize

A bunch of plug & chug stuff including EL, UEL, Sharpe, Trynor, information ratio, find put delta given N(1), umm.....CAPM, CAPM assumptions, bionomial interest rate tree (1 period). Speaking of that, I almost forgot I had to discount it BACK to the PV. Because there was an answer for the undiscounted value. Sneaky sneaky.

 

[Rohanm94]

What would a safe score for passing be?

 

[Detective]

@VWJETTY

For Poisson I think choices were like, 7%,10%, 40%, 80% (rounded). I was thankful for that one they made choices spread out because I was not looking forward to using Taylor series expansion for e^(-lambda * t) to approximate w/o calculator.

If I recall correctly, lambda = 2, t = 2, and question asked about probability n=6 exactly. I think I put either 10% or 40% (not sure why I would even think 40% if lambda = 2). It does appear the answer is around 10% now that I used a computing engine:

https://www.wolframalpha.com/input/?i=(2*2)^6+*+exp(-2*2)/(6!)

For the 99% VaR, you're right you take an average of the lowest and second lowest (that's what I did). I wasn't sure if was lowest/2nd lowest OR 2nd lowest/3rd lowest though, but it seemed 2nd/3rd lowest average wasn't even a choice.

FRA is essentially a single period IRS. I.e. IRS is a summation of FRAs.

For the teacher question you mean like last question on the exam (i.e. number 100) or the last question you left blank? That question was near the middle for me if I recall, so wonder if they scramble the exam questions across centers / where you sit (probably do). I don't remember the choices or answer though.

For the 1 tree binomial tree, this might be a stretch, but do you remember if you put 0.81 or 0.83 (if those were even choices). Basically, I didn't have a functioning calculator at that point, and I wasn't able to pull out enough tricks to figure out if it would be 0.81 or 0.83, I went with 0.81.

 

[shantharam1]

I felt that the actual was tougher than the GARP mock solely due to more number of qualitative questions asked, which added more confusion. I could solve 30 around questions (with 85% accuracy) in the first one hour and from then on was going at 50% accuracy.

@VWJETTY Yes, ES would be the average of the last two.

Can you guys clarify on the 1-step binomial tree question, it was given that Uptick factor was 8% and Down tick factor was 4%. So what would the value of D be for calculating probabilities? D=1/1.08 or D=0.96. I used 0.96 and got the answer, but I am confused as hell.

Also, I remember there were sums on Factor Model(Arbitrage Pricing Theory), one question regarding seasonality, one based on confidence intervals of regression([0.79,IDK] Upper and Lower bounds, cannot reject the hypothesis), conversion of 1% VaR to 5% VaR, calculating long run average variance, protective put, etc.

 

[VWJETTY]

Yes, I think for the poission I put 10%. Fairly certain on that one.

Cool on the ES one. Because I too took the other average and it wasn't even an answer so I think we may be good on that.

Ahhh....I forgot what I put for the FRA but yeah, that makes sense.

The teacher question was my last question. So yeah I think they do scramble the questions around.

the 1 tree binomial is 0.81 I think. Because the 0.83 one was without discount to PV.

For the put-call parity one, I think I said short the call and short the stock.

 

[VWJETTY]

for that one question about ERM, I said it would allocate risk capital. The other 3 answers seemed off. Not sure if you know what I am talking about but it was one of the qualitative questions.

 

[alexvaag12]

Very easy exam... Honestly it was way behind my expectations. BT helps prepare in details that make the exam look like a joke. Only tricky theoretical questions, but problems were just straight forward.

 

[Detective]

for that one question about ERM, I said it would allocate risk capital. The other 3 answers seemed off. Not sure if you know what I am talking about but it was one of the qualitative questions.

Yeah, I know what you're talking about, but the qualitative questions are fuzzy to me since most of my time was spent on quantitative ones. Even though some of them were tricky I do recall them being easier than expected b/c

-They weren't that "wordy". Reading previous exam threads I was prepared to sift through paragraphs of text to get to the point, but I thought they were concise and to the point.

-They asked "which one is correct?" rather than "which one is NOT correct", and I think that is personally easier.

For the specific question you're talking about, I, unfortunately, do not remember the choices or what I put, but it makes you feel any better I vaguely do recall gravitating towards risk capital. I remember only two qualitative questions somewhat well b/c I think I got them wrong.

(1) Asked which is correct related to MBS with choices?
-2 Choices that I think were wrong
-"OAS" being a spread to interest rate paths (but the way it was worded seemed wrong)
-The prepayment option of MBS is attractive when PV of scheduled principal payments is less than the outstanding principal

(2) If you think stock value will double what strategy is best?
-Protective Put
-Covered Call
-Bull spread of call options
-Bear spread of call options (or might have said put options)

 

[Detective]

I felt that the actual was tougher than the GARP mock solely due to more number of qualitative questions asked, which added more confusion. I could solve 30 around questions (with 85% accuracy) in the first one hour and from then on was going at 50% accuracy.

@VWJETTY Yes, ES would be the average of the last two.

Can you guys clarify on the 1-step binomial tree question, it was given that Uptick factor was 8% and Down tick factor was 4%. So what would the value of D be for calculating probabilities? D=1/1.08 or D=0.96. I used 0.96 and got the answer, but I am confused as hell.

Also, I remember there were sums on Factor Model(Arbitrage Pricing Theory), one question regarding seasonality, one based on confidence intervals of regression([0.79,IDK] Upper and Lower bounds, cannot reject the hypothesis), conversion of 1% VaR to 5% VaR, calculating long run average variance, protective put, etc.

D = 0.96, basically 1-4%. Per VWJ, answer was around 0.81. That's what I thought it should be too.

Seasonality I went with the first choice which said something to the effect of that some days are more volatile, though the choice saying that "you cannot see seasonality effect if you have only 1 year or less of data" seemed tempting, I wasn't fully convinced that was necessarily true; however, it sounds more right now than the choice I put now if I think back to it...

For CI I think there were potentially two questions, I do remember saying "reject null" though, but I could be remembering wrong.

I don't remember conversion of 1% VaR to 5% VaR, but I could be misremembering or missed that part of the problem.

For the long-run average variance in GARCH(1,1), it was simply, omega/(1 - alpha - beta), very common test question.

 

[shantharam1]

D = 0.96, basically 1-4%. Per VWJ, answer was around 0.81. That's what I thought it should be too.

Seasonality I went with the first choice which said something to the effect of that some days are more volatile, though the choice saying that "you cannot see seasonality effect if you have only 1 year or less of data" seemed tempting, I wasn't fully convinced that was necessarily true; however, it sounds more right now than the choice I put now if I think back to it...

For CI I think there were potentially two questions, I do remember saying "reject null" though, but I could be remembering wrong.

I don't remember conversion of 1% VaR to 5% VaR, but I could be misremembering or missed that part of the problem.

For the long-run average variance in GARCH(1,1), it was simply, omega/(1 - alpha - beta), very common test question.

Great, thanks.
Also, there was one question on the poisson put (Option C, AFAIR).

There was a question on Unexpected Loss and I remember marking 22,760 for that, was that right?

 

[Detective]

Great, thanks.
Also, there was one question on the poisson put (Option C, AFAIR).

There was a question on Unexpected Loss and I remember marking 22,760 for that, was that right?

Yeah, I went w/ Poison Put as the answer on the one you're talking about.

For the Unexpected Loss, my calculator was not working, so I guessed and do not remember.

 

[Rohanm94]

Great, thanks.
Also, there was one question on the poisson put (Option C, AFAIR).

There was a question on Unexpected Loss and I remember marking 22,760 for that, was that right?

Yes, remember marking 22760.
Poison put regarding a question on bonds for M&A? Probably yes.

 

[shantharam1]

Yes, remember marking 22760.
Poison put regarding a question on bonds for M&A? Probably yes.

Yes the one on M&A and regarding the passing score, my guess is at 55-60. Others who think otherwise, please chip in.

 

[VWJETTY]

in regards to the previous chain thread....

(1) Asked which is correct related to MBS with choices?
-2 Choices that I think were wrong
-"OAS" being a spread to interest rate paths (but the way it was worded seemed wrong)
-The prepayment option of MBS is attractive when PV of scheduled principal payments is less than the outstanding principal

Yes, these two were wrong. But i forgot the answer I put down but i know for a fact i didn't select these two.

(2) If you think stock value will double what strategy is best?
-Protective Put
-Covered Call
-Bull spread of call options
-Bear spread of call options (or might have said put options)

I said bull spread of call options.

Also, yes for that other question, I said poison pill as well.

for the 1-day 99% to 10-day 95% VAR. You had to...

1) divide by 2.33
2) times by 1.65
3) times by sqrt 10

that's what I did at least.

 

[Detective]

Yes the one on M&A and regarding the passing score, my guess is at 55-60. Others who think otherwise, please chip in.

I feel 55-60 is too low for pass. I was thinking high 60/low 70 for a pass.

Maybe I am biased, and I am not sure what you used to prepare for the exam (if any third party material at all). In an ideal world, I would've read all the source material, written my own notes, and done end of chapter problems from the source directly without any third party prep.

Real life constraints made this more of an idealization and after doing some research on third party prep I purchased BT product ~2 months before the exam. I am very satisfied with the purchase, and as other subscribers here have hinted that after BT prep, the actual exam comparatively felt like a cake walk. Again, since I did not use any other products, I cannot really comment if their product would have made me feel prepared.

Even if I do not pass this time due to my own fault (calculator issue & lack of prep on certain topics - 3&4). I am fairly confident now given what I have seen if I use BT prep, and earnestly prepare this time for around a month, I can get 90+ on the L1 exam.

 

[subha107]

There was a question on Bayes Theorem could not get the correct answer. Anyone got correct match ?

 

[VWJETTY]

I'm not sure what the marginal passing score is for this exam though. Does anyone have a rough idea how many questions you need to pass this exam?

There was also a ethics question about confidentiality. I said you need consent before you can use confidential information.