Exam Feedback November 2021 Part 1 Exam Feedback

[Nicole Seaman]

We hope that everyone did well on the FRM Part 1 exams over the past two weeks! We would love to hear any feedback that you have about the exam, especially any helpful information that you can provide about the new CBT format. How did the exam go? Did you encounter unexpected questions? How was the experience with the COVID guidelines? Thank you in advance for any feedback you can provide! (Note: The above posts were shared before I had the chance to create this thread)

 

[sulemanms202]

i cant discuss the specifcs but i can tell you paper was very hard even though it was very qualitative. i could hardly attempt 65 mcqs in 3 hours and did 25 just on guess work

 

[Garbanzo]

Hi everyone,

I just had the exam yesterday and feel the need to dump this here today already to clean up my mind and while it's still fresh. The last exam day according to GARP was yesterday, November 26, too, so it should be fine. From what I read here previously, it's ok to share the specifics of the exam (but I remember in the pre-exam declaration it said sth about agreeing to not reproducing the exam even from memory - just don't know if that means only until the exam window is over or forever - judging from the many detailed exam feedbacks on this forum, I guess it is ok), but if it is too detailed and violates the rules, please let me know.

First, the previous exam feedback threads (May/July 2021) were quite useful, so thanks to all who contributed. The GARP practice exam was useful as well - slightly easier, but still filled a gap here and there.

After finishing the exam, I couldn’t see the score when clicking on the results and back button.

I think it was a fair exam, I liked that it didn’t test extremely many formulas, and it was mainly about applying logic and familiarity with the concepts. There were a few wtf questions, like always.

Here we go:

• Jarque-Bera – if you understand the formula, you’re good (the statistic depends on sample size, null tests that skewness is 0 and kurtosis 3 (data is normally distributed), chi-square distribution with 2 degrees of freedom)
• Kendall’s tau/Spearman correlation (they were just answer choices, but you need to know the basic idea/use cases) they are not in the study notes, so you may google or check the book. I read these two pages/videos:
• https://www.statisticshowto.com/kendalls-tau/
• https://www.statisticshowto.com/pro...earman-rank-correlation-definition-calculate/
• 99% ES. Calculate given a uniform distribution from -10 to 12. Don’t make the mistake of choosing the VaR level instead of going beyond the 99 and calculating an average of that – which is equivalent to 99.5% VaR. (the answer is (-1) * 22 * 0.995 + 12 = -9.89)
• Dutch auction – very easy but you need to know that the price is the lowest one that ensures allocation. And how many shares each participant gets (highest bidders get full allocation and the lowest successful bidder that decided the price – most likely only partial)
• CAPM – usual Beta, expected return, Sharpe calculations etc. But one weird question – determine systematic risk and unit price of market risk, when given market sigma (that’s an answer to the systematic risk part right there) and portfolio sigma and correlation and sth else.
• SML vs CML – how does the slope change with change in risk-free rate, when to use which (portfolio vs individual security), what’s the x-axis risk.
• APT – with a very weird twist that I couldn’t figure out. It looked like they didn’t give enough inputs to answer, and the options looked like they were just a trick. They gave inputs for “sensitivities”, “risk premiums”, “risk-free rate” and I think Beta or correlation. And you should determine the error term in APT, or beta or what’s the input to the APT in terms of one of the factors.
• Information ratio (don’t fall for the trap of using standard deviation of the portfolio return but use the active return standard deviation), also know that obviously higher information ratio is better.
• Treynor ratio (uses Beta not standard deviation)
• Calculate confidence intervals and determine if null rejected
• Determine t-test degrees of freedom and properly choose a one sided or two-sided quantile for a given significance/confidence level
• Formulation of null and alternative hypotheses – this was weird again as none seemed to be correct. The null must always have an equal sign and AFAIK the alternative must cover the rest (i.e., if null is an equal, alt is not equal; if null is less than or equal, alt is greater than, and vice-versa). The question was about a one-tailed test, yet the only option where null and alt hypothesis covered the whole distribution was a null equal to zero and an alt not equal to zero.
• EWMA – be careful about which term the (lambda) and (1 – lambda) weights correspond to (calculate one-step ahead forecast or current value given the inputs. Another question was what do we need to calculate it except the AR(1) term –> lagged squared return). Understand the basic idea of the decay in weight the further the observation goes to history. Also that its long run is 0.
• Equally weighted historical volatility – what’s the limitation (if a large point long in history drops off at some point in time, it may change the model value today significantly)
• ARMA (1,1) mean reverting level - (I think should be the same as AR(1) so intercept/(1-AR_coefficient))
• A few questions on Monte Carlo Simulation/Historical simulation (“it’s best practice to scale the shocks equally for all variables” – I don’t think so )/Stress testing - advantages, limitations, best practices and theory
• A few questions on Exchange vs OTC comparisons, some were kind of detailed, collateralization vs margins, comparative size of the markets, what instrument has highest notional (both are IR derivatives, but they tried to trip you up by saying Exchanges have forex as the most popular), are OTCs mostly bilateral?
• Effect of transformation and scale on mean/variance/skewness/kurtosis – be careful here, the only correct was that kurtosis is invariant. The skewness changes sign if the scaling factor is negative, but the answer choice was “-b” (i.e. including the scalar) which is incorrect
• Conditional probability/Bayes rule with a twist – need to transform the probability matrix a bit first; a basic question on conditional probability – but need to calculate the joint probability by yourself when given the unconditionals and union probabilities (so the joint is sum of unconditionals minus the union)
• Hazard rate – this was a bit tough for me because I haven’t done many examples; be clear on cumulative/conditional/unconditional default/survival probabilities
• Lots of questions on ERM/governance (quite detailed – role of BoD vs Senior management in stress testing – whose primary responsibility is adequate stress testing staff compensation and whose ensuring qualified people do the stress testing, best practices etc)
• Risk culture – I think 2 – 3 questions. Mostly strange unfamiliar statements (I chose that Machine learning can provide good indicators of when issues in risk culture may lead to losses – it seemed the most reasonable)
• 2 cases – Metallgesellschaft and Orange County – lessons learned (careful about your hedging strategy + don't use derivatives you don't understand)
• Coming out of GFC – what changed (sth about ERM, also how the OTC market changed)
• CRO compensation and reporting to CEO directly
• Delta-normal approach to VaR
• Poisson distribution – calculate volatility in dollar terms given PD and security price, they mentioned the portfolio consisted of 25 securities but asked about the volatility of a single one
• Bull spread with two calls – final payout (NOT profit)
• A few BSM questions
• One step binomial tree Am call (no difference from a European really) – but the question had two parts – calculate value at the first node AND determine the delta using the final two nodes – I forgot to use the formula ((up_value – down_value)/(up_price – down_price)), but because the option was in the money, I just chose the delta higher than 0.5 - there were only two distinct choices of ~0.23 or ~0.73. However, now that I am thinking back about it and the spot/strike/up/downs, I think the correct answer was actually the lower delta.
• Downward sloping spot curve – how does the rank order with par curve and forward curve look
• Covered interest rate parity vs Uncovered – a bit twisted answer choices, not the basic stuff
• Country risk and sovereign credit rating, easy
• Can a country default on domestic debt? + foreign, etc. What’s the effect of printing more money (also on the exchange rate)
• Ops risk – Basel II BIA vs SA vs AMA – even though I knew all, this was tricky. I think the correct answer was that a bank that used to use AMA had to have a regulatory “permission” (strong justification?) to go back to BIA (but this I read two and a half years ago when I prepared for my job interview and not sure xD; the other options were if BIA calculates capital for each op risk type; and something wrong about SA). Nothing about the current ops risk SA or SMA
• Synthetic put using a stock index
• Delta-hedging – was a bit tricky – the delta was negative, so we need a long futures position in the amount of delta
• Something like calculate forward interest rate given interest rate futures and volatility – didn’t know how to do this, I just remembered the basic formula of fwd = fts – 0.5*variance*T1*T2 but there were no dates AFAIR and the underlying was an interest rate
• 4 options (2 American, 2 European) on 2 similar stocks with the only difference being a dividend paid or not. What’s the rank order of their prices
• Gaussian White Noise and iid and normal distribution (understand how these relate with the types of white noise). Tricky choices, I think it was about choosing an incorrect one.
• Covariance stationarity assumptions – which of the 4 options is a must – I think there was a statement that autocorrelation (dependence) is OK for the included variables BUT the error term autocovariance must depend only on the time displacement
• Basic OLS assumptions – but weird question/answers – two options seemed correct – no large outliers in the explanatory variable, and random draws of X&Y (but I think this option was worded improperly)
• Calculating expected loss given recovery rate, PD, EAD, and WCDR (I think we were supposed to use the WCDR – PD adjustment but I don’t remember how the question was worded. In the study notes it said this is an IRB approach, but the question didn’t mention it.
• Conditional dependence/independence
• A few “case studies” where people want to test some sort of a hypothesis
• Understand the difference between the standard deviation of a sample and the standard error (se = sd/sqrt(N)), but know when/how to apply
• Gaussian copula basic idea
• Convexity calculation – remember that “delta y” is only the move from the original level to one of the sensitivity levels (e.g. 3.95%; 4.00%; 4.05% – delta y is 0.0005 in decimal not 0.001; they give answers that you would get using either of those)
• A coupon bond – given beginning price and end price, price change due to carry roll + due to rate change; then you know the coupon is 2%. Calculate how much the bond price changed due to spread change. A simple equation with one missing variable, but you should ignore the coupon.
• Multicollinearity
• Regression with dummy variables for each month except December – which coefficient measures the December effect? (intercept)
• Omitted variable bias/including irrelevant variables/ bias-variance trade-off
• Forward buckets vs Key rates vs Partials – strange answers (sum of key rates is the shift in YTM? (my choice) However, the mock exam answer key says that sum of key rates is a shift in the par curve, so go figure ; it is stated that we should use partial 01s when using swap curve (which is the par curve so then how can a sum of key rates be a shift in par curve?)), how do the shifts influence something (don’t remember).
• Implied volatility and VIX (they were listed as two different answer choices) – are they forward-looking, backward-looking, how are they calculated (current prices or historical prices)?
• Very simple question on economic capital – a chart with the loss distribution and a few lines (vertical mean, mode, 99% VaR, horizontal line from mean to 99% VaR (that’s the EC))
• Understand that EC is only for unexpected losses
• Simple question on expected credit loss of a portfolio (when does it decrease? -> when recovery rate goes up)
• Property and Casualty combined ratio – what variables are inputs – but it was tricky because the correct option only had two out of the three (expenses and claims, without the premium revenue), but all the other options had the inputs to the operating ratio (dividends, investment income AND premiums revenue). So, you need to be familiar with all.
• Trustee in indenture – role, but again that tricky stuff about covenants (check, enforce, formulate) and what the trustee can/has to/may or may not etc. Not a clear answer to choose.
• Covenants for high quality issues are fewer (I think).
• Hedge fund strategy that capitalizes on a down market BUT still has the potential to gain on some well-performing companies (long-short; vs managed futures, convertible bond arb, dedicated short)
• Futures arbitrage (given spot, interest rate, two futures and interest rate, easy)
• Number of contracts to hedge a position + hedge ratio (understand it’s the beta of a ‘regression of the asset returns to hedge on the returns of the hedging instrument’. ALSO understand that ‘Beta = correlation * (sigma asset/sigma hedging instrument)’ be careful about the order of the cross-volatility inputs. There was a cross-hedge question (basically the same logic as in the study notes – jet fuel hedge with heating oil futures contracts)
• Adjusting the beta of a portfolio using futures (how many contracts)
• Understand what’s a contango and backwardation (relationship between spot/near/long dated futures prices) and when do we lose/gain on long/short futures positions (and don’t confuse with normal contango/backwardation that refers to the relationship between the unobserved expected future spot vs futures prices – but understand this as well + what’s their relationship when the required rate of return is higher/lower/equal to the risk-free rate)
• Basket options, Asian options – basic logic
• How is the storage cost generally thought of – adjustment to income, convenience yield, interest rate, or dividend yield. (I chose interest rate, but maybe it’s convenience yield like in the lease rate calculation? Annoying question, like what’s the point…)
• We shorted 300 stocks, there are two dividend payouts before we close out the position. You’re given initial price, final price, dividend size, transaction costs (per 100 shares) – what’s the net profit. (Remember, when shorting, you don’t own the stock – need to pay the dividend to whom you borrowed it from). The trick was also that they worded it so that it seemed like there was only one dividend (“the company announced it will pay it on this date”) but another part mentioned that they determined the exact size of the dividend for the next year and it will be paid each quarter end. The time horizon included two quarter ends – so there were two dividends. Also remember that to close out a short position you need to pay the transaction cost again. So, there would be a x2 in the transaction costs (it was given per 100 shares, so you need that number x6 (300 shorted, 300 bought back))
• Basic question on a forward rate calculation given a spot rates table.
• Callable bonds, tender offer – what the company can/cannot do to call the bonds, are all corporate bonds callable? (Obviously not)
• Understand what’s a plain vanilla derivative and distinguish from exotic options by being given 4 instruments characteristics.
• Credit spread relationship with interest rates, how is it set, what does it contain
• Mortgage-backed securities – understand CMO basic logic, what’s a TBA, what’s a pass-through (vs interest/principal only); prepayment – who benefits?
• Calculate treasury bill cash price from quoted price (remember to use 30/360 convention, the days are given; another answer choice was also for when you would incorrectly use 30/365 convention)
• Is there something like a fixed-for-floating COMMODITY swap? (I think not, but it was one of the choices for a question on the best hedge against commodity prices falling – I chose put options even though they cost money so you could argue they are potentially not THE best, but there was no simple forward as an answer choice, the other two were some nonsense)
• Understand a covered call (what the investor’s view on price is – essentially his view is that it will rise but not really cross the strike, unless he is ready to sell for that price. It’s a nice income enhancement strategy when we’re ok to sell for the strike target price)
• EAD of bonds and derivatives (do they both take the current price as the EAD? No, derivatives need more complex modelling)
• What does credit rating agency’s outlook and watchlist mean (outlook is medium term and watchlist up to 3 months, but I think there was some other characteristic too)
• Companies internal credit scoring – what does it use (quant methods or mix of qualitative/quant too?), is it inferior to credit ratings?
• Point-in-time and through-the-cycle – basic ideas about stability, when to use which given some case facts (I think mostly long vs short-term horizon usage).
• Using external database for loss severity is good (with adjustment for size), but not good for frequency
• Should a company eliminate all operational risks that it can eliminate? (best practices question with some other choices)
• Stress testing vs VaR
• Bullet vs barbell portfolio with the same duration, which one will gain more/lose less in parallel up and down shift (barbell is better because it has more convexity)
• Calculate IQR when having 12 observations. Another question was the same but when having 9 observations – here it’s a bit tricky to properly average the numbers.
• Calculate a 2-year term life insurance premium for a 70-year-old woman if it’s assumed she will die in the middle of “a” year (I think you need to discount 0.5 and 1.5 and get the expected value given the probabilities she would die then and choose the bigger value)
• Antithetic variates and control variate techniques to reduce Monte Carlo sampling error – you need to understand these techniques kind of well because the answers were worded quite ambiguously where just a single word could mean it’s incorrect.
• GARP Code of conduct – 2/3 in total, very ambiguous like all ethics questions. Not sure how we are supposed to answer them, the code of conduct is very short, yet the question “cases” were going into some details you can’t figure out just from reading the code. And the options seemed to either all go against or be in line with the conduct. (is it in violation to appoint an external party to write your company’s code of conduct? – I seem to remember sth like this from CFA but not sure) (Another was if it is in line with the CC if we allow employees to take religious holidays (obviously); restrict senior management from being on BoD of other companies we do business with (yeah, I would lean towards this, but then maybe if you disclose this it’s ok?), etc.
• one more - I think it is true that management can be incentivized to long term value creation (or some sort of conflict of interest issue) if their compensation would be decreased if the capital dropped below minimum regulatory capital. It was the only reasonable choice from what I remember.

This adds up to around 95 questions that I remember.


All in all, to get through the exam I think one needs good familiarity with a lot of the concepts because the answers are sometimes just ‘one-word-wrong’ and usually quite twisted, the inter-relationships are important. + formulas memorized to the details, because again, the answers differ so that when you switch around the parameters (like in the EWMA) you will get one of the incorrect choices. Same with the convexity formula – you need to know the “delta y” is not the full range of y-changes up and down, but just the one half of the range. Same with the correct “cross-volatility” order in the formula for an instrument’s Beta (= correlation * sigma asset/sigma benchmark). It is so frustrating to know the concept quite well but get tripped up on points like these.

 

[gzeidan]

I don't have Garbanzo's memory bank but interestingly I had quite a few different questions so it seems we had very different papers. I would tend to agree that the exam was relatively fair but it was definitely more theoretical (maybe 65%). I did have some time management issues but that was because i really took my time in the first 50 Qs and then had to speed up. I would also add that although i did the topic drills, i didn't do any mock exams as i just didn't have the prep time. The main thing I noticed is that the exam relied more on nuances and points you pick up from reading and re-reading rather any detailed analysis. Otherwise there were the standard topics to be expected e.g. EWMA, bayes theorem, Monte Carlo, Code of Conduct etc.

 

[sulemanms202]

easy mcq (so easy didn’t even require much calculations)

• Blue (selected one which blue is one which has least variance amongst unbiased ones)
• Dutch Auction (300,000 shares answer)
• MVHE
• 4 conditional probabilities (AUB, 28% -- 2 answers I remember)
• 2 EWM (0.33% and lamda smoothing – answers)
• Portfolio manager shares shorted (answer 720$)
• Bull spread (answer 2.39$)
• Economic capital ( I made mess of that one ) and EL calculations (easy PD LGD EA)
• 4 from credit ratings
• ERM 3 MCQs
• Risk appetite
• Beta and CAPM
• Bernoulli distribution and loan portfolio problem (3000 CAD answer)
• Very easy CLT and sampling distribution MCQ (didn’t even require full reading!)
• Another easy one on PD and LGD relationship
• MGM case (I answered Funding is required for stack and roll)
• GFC ( I answered: they kept CDOs on their Balance sheets)
• MBS
• MCS : large number of replications required
• ES and uniform distribution (9.89 answer!)

Confusing ones

• Backwardation vs contango
• Quantitative ERM (I wasted almost 5 minutes on that!)
• Stupid code of ethics problem (almost all looked true)
• Multiple linear regression problem (I got stuck with it and answer is certainly not dummy variables option!)
• 2 MCQs - Corporate bonds trustee (almost second option looked right!)
• Option bounds (so confusing in exam conditions!)
• Operational risk AMA
• 4 ARMA MCQS

I got wrong or made educated guesses!

• Delta hedge
• Comparision of bonds
• Yield on bonds
• Swap valuation
• Put option on swap ( I selected first option disregarding second option – credit risk which was right)
• Very easy CAPM mcq theoretical – completely made mess of that
• Sharpe and treynot ( I got sharpe 2.33 right and wasn’t getting treynor so selected 3.33 I think!)
• INR bonds comparision
• Very confusing confidence interval problem
• Binomial model
• Term structure and forward rates
• Forward rates calculation ( I still cant understand what problem was about)

 

[sulemanms202]

in contango/backwardation and convenience yield MCQs... trust me it was very difficult to identify facts..

 

[sulemanms202]

More easy ones: ( and i got them right!)
implied volatility (liquidity based thus requires historical data)
market risk how not involved in OTC (no matched books in OTC! in other words no novation)
role of collateralization and how to value collateral (certainly not based on average daily prices!)
Margins, OTC and exchanges (draw comparison!)

 

[sulemanms202]

stress testing and tolerance limits ( i got that badly wrong!)
scenario analysis (very easy and doable)

on the whole BT material was very good only problem i have is with feedback. See in online education there is no live feedback and that was THE problem for me. I think had i had feedback on my practice issues i would have easily scores 80+ but now i am struggling to get 50

by my calculation i will end up with around 42 to 45

 

[sulemanms202]

2 easy ones that i can recall now.
mutual funds (3 obvious wrong choices!) right one was payment of commission to broker for getting clients( something along the lines of, cant remember exactly!)
breakeven premium (heart aches i did that wrong)!

 

[vinebills]

exam was too qualitative is what i felt. You had to know the material really well, else you would be lost. Overall a good exam- will i pass or not - well I would give myself a 50% probability. BT was very helpful. Thanks to David and Nicole!

 

[sulemanms202]

not much clarity in choices, it was hard to relate back to what i had studied. Also, there was not any balance in the way areas were covered. In the quant section my paper had just 2 mcqs on hypothesis while ARMA had around 5 mcqs and of those 2 mcqs one was rather difficult to understand. But BT material was really a top notch standard, has prepared me very well and i cant thank them enough.

 

[yisuho97]

It's been a while since I took the exam (11/19 Chicago), but it was hardly quantitative. I remember going over BT practice questions and thinking to myself there are too many formulas I have to remember, but it felt like 30% quantitative and 70% qualitative. Calculations were easy thanks to BT, but I had a hard time going through qualitative part + I'm new to financial area. Felt like balance among different topics was missing too. Had little statistics while much was asked about risk theories. Score was unavailable at the end of exam. Fingers crossed!

 

[Chris]

I took mine on 11/17 Toronto, and I'd say close to 75% qual, 25% quant. It feels like I guessed close to 35% of the questions. No questions on PCP The sample exam provided by GARP isn't close at all to the FRM content. Good luck guys

 

[vinebills]

any idea when we get the results?

 

[sulemanms202]

results are out i did not pass

images

can anybody else confirm this?!!!!!!

Change your PC time clock to jan 17, 2022 an restart and log in to your account.

 

[Nicole Seaman]

Change your PC time clock to jan 17, 2022 an restart and log in to your account.

If you look back at the feedback threads from the past, you will see that "tricking" the system does not always provide the correct results. I have seen people change their time clock and show pass results when they actually failed and vice versa. As difficult as it is to wait for the results, I encourage everyone to wait until GARP releases the official results so they are accurate.

 

[sulemanms202]

idk nicole... this has provided me relief..now i can focus towards my work!

 

[vinebills]

thanks Nicole & Sulemanms - now I too can focus on work

 

[sulemanms202]

lol...smiles!

 

[diofar]

If you look back at the feedback threads from the past, you will see that "tricking" the system does not always provide the correct results. I have seen people change their time clock and show pass results when they actually failed and vice versa. As difficult as it is to wait for the results, I encourage everyone to wait until GARP releases the official results so they are accurate.

I logged in garp.org about an hour ago and results were released and I passed it! But now it seems they are not available anymore! I hope they will not change the result to fail!