FRM Part II Exam

By AnalystPrep

Questions with Answers - Market Risk Measurement and

Management

Last Updated: Feb 11, 2023

1
©2023 AnalystPrep “This document is protected by International copyright laws. Reproduction and/or distribution of this document is

prohibited. Infringers will be prosecuted in their local jurisdictions. ”

 Table of Contents

Estimating Market Risk Measures: An Introduction and

63 - 3
Overview
64 - Non-parametric Approaches 23
65 - Parametric Approaches (II): Extreme Value 42
66 - Backtesting VaR 64
67 - VaR Mapping 81
Messages from the Academic Literature on Risk
68 - 102
Measurement for the Trading Book
69 - Correlation Basics: Definitions, Applications, and Terminology 111
Empirical Properties of Correlation: How Do Correlations
70 - 133
Behave in the Real World?
71 - Financial Correlation Modeling—Bottom-Up Approaches 142
72 - Empirical Approaches to Risk Metrics and Hedging 149
73 - The Science of Term Structure Models 165
The Evolution of Short Rates and the Shape of the Term
74 - 181
Structure
75 - The Art of Term Structure Models: Drift 190
76 - The Art of Term Structure Models: Volatility and Distribution 214
77 - Volatility Smiles 231
78 - Fundamental Review of the Trading Book 244

2
© 2014-2023 AnalystPrep.
 Reading 63: Estimating Market Risk Measures: An Introduction and
Overview

Q.1471 During a job interview for the role of assistant financial risk manager, Jacob Lee was asked to
describe the advantages of geometric returns over arithmetic returns. He mentioned the following
point:
I. T he geometric mean takes into account the compounding that occurs from period to period.
II. In the world of finance, the arithmetic mean is usually an appropriate method for calculating multi-
period average returns if the one-period returns are volative.

Which of the above-mentioned points is/are correct?

A. Only Point I is correct

B. Only Point II is correct

C. Both I and II are correct

D. Neither I nor II is correct

T he correct answer is A.

Point I is correct. T he geometric mean differs from the arithmetic average, or arithmetic mean, in
how it's calculated because it takes into account the compounding that occurs from period to period.
Because of this, investors usually consider the geometric mean a more accurate measure of returns
than the arithmetic mean.
Point II is incorrect. For volatile numbers, the geometric average provides a far more accurate
measurement of the true return by taking into account period-over-period compounding. For
example, if your portfolio returns each year were 90%, 10%, 20%, 30%, and -90%, the arithmetic
average would be 12%. However, the geometric average annual return would be -20.08%, which is a
lot more accurate.

3
© 2014-2023 AnalystPrep.
Q.1472 If the arithmetic returns are 0.10, which of the following equations gives the corresponding
geometric returns?

A. R t = ln(0.10)

B. R t = ln(1.10)

C. R t = ln(1.90)

D. R t = ln(0.90)

T he correct answer is B.

If we have the arithmetic returns, rt , we can find the corresponding geometric returns from this
equation: R t = ln(1 + rt ). In this case, R t = ln(1 + 0.1).

Q.1474 VaR is a quantile that demarcates the tail region and non-tail region. Which metric is used to
calculate the size of the tail?

A. Standard deviation

B. Confidence level

C. Number of observations

D. Variance

T he correct answer is B.

In theory, the VaR is the quantile that separates the tail region from the non-tail region, where the
size of the tail is determined by the confidence level. For example, suppose we have 1000 loss
observations and are interested in the VaR at the 99% confidence level. T his means we would have a
1% tail, and 10 observations in the tail. T he VaR would be the 11th highest loss observation.

4
© 2014-2023 AnalystPrep.
Q.1475 If Profit/Losses (P/L) are distributed normally with standard deviation 18 and mean 12, then
what is the value of the corresponding VaR using a 95% confidence interval?

A. 9.87

B. 17.61

C. 13.956

D. -13.956

T he correct answer is B.

If P/L over some period are normally distributed with mean 12 and standard deviation 18, then the

95% VaR is given by:

αV aR = −μP/L + σP /L Zα
= −12 + 18z0. 95
= −12 + 18 × 1.645 = 17.61

Q.1476 If the natural log of X is distributed normally, then the random variable is:

A. Normally distributed

B. Lognormally distributed

C. Seminormally distributed

D. Equally distributed

T he correct answer is B.

A random X variable is said to be lognormally distributed if the natural log of X is normally distributed.

5
© 2014-2023 AnalystPrep.
Q.1477 Which of the following statements regarding the estimation of the expected shortfall (ES) is
false?

A. T he ES can be estimated as an average-tail-VaR because it's actually the probability-

weighted average of tail losses.

B. T he ES can be estimated with the help of a 'closed-form' solution.

C. A 'closed-form' solution is more preferable than the average-tail-VaR because the formula
and its application is widespread in all parametric distributions.

D. T he average-tail-VaR is more preferable method to estimate the Expected Shortfall (ES) as

it has higher applicability and is easier to implement.

T he correct answer is C.

A 'closed-form' solution can be a useful way of estimating the ES, but ES formulas seem to be known
only for a limited number of parametric distributions. T he 'average-tail-VaR' is comparably easier to
use and can be applied to most ESs encountered in practice.

6
© 2014-2023 AnalystPrep.
Q.1479 A generally coherent risk measure tends to involve increasingly sophisticated weighting
functions. Which of the following is a suitable replacement for the equal weights in the 'average VaR'
to estimate any risk measure?

A. Average weights

B. Exponential weights

C. Weights appropriate to risk measure being estimated

D. None of the above.

T he correct answer is C.

It is possible to estimate coherent risk measures by manipulating the “average VaR” method. When

measuring risks, it is important to assign weights taking into account the specific risk measure being

estimated. T hese weights assigned, helps to estimate coherent risk measure, and are a suitable

replacement for the equal weights in the 'average VaR'.

A coherent risk measure is a weighted average of the quantiles (denoted by qp of the loss

distribution):

1
M∅ = ∫ ∅ (p) qpdp
0

where the weighting function ∅(p) is specified by the user, depending on their risk aversion. T he ES

gives all tail-loss quantiles an equal weight of [ (1–1 l) ] and other quantiles a weight of 0. T hus the ES is
c

a special case of M∅ .

Under the more general coherent risk measure, the entire distribution is divided into equal

probability slices weighted by the more general risk aversion (weighting) function.

7
© 2014-2023 AnalystPrep.
Q.1480 T he key to estimating coherent risk measures lies in the:

A. Ability to assign weights accurately.

B. Ability to calculate exponential value accurately.

C. Ability to estimate quantiles.

D. Ability to approximate risk exposure.

T he correct answer is C.

If quantiles are estimated appropriately, then the coherent risk measures can be easily calculated by
finding the average.

Q.1481 T he precision of a risk measure estimate is evaluated by means of the corresponding

standard error(s). T he quantile (VaR) standard error depends on which of the following?

A. f (q), sample size n and p

B. p, standard error s and variance of q

C. sample size n, p and square root of error

D. variance of q, sample size n and f (q)

T he correct answer is A.

According to Kendall and Stuart, the standard error falls as sample size n rises; it rises as probabilities
become extreme and more toward the tails; and it also depends on the probability density function
f (q).

8
© 2014-2023 AnalystPrep.
Q.1482 A portfolio has a beginning period value of $200. T he arithmetic returns follow a normal
distribution with a mean of 5% and a standard deviation of 10%. Calculate VaR at both the 95% and
99% confidence levels, respectively:

A. $23, $36.6

B. $43, $56.6

C. $1.65, $2.33

D. $23, $43

T he correct answer is A.

Using a parametric estimation approach,

VAR(α%) = (−μr + σr × Zα ) × P t−1

Where:

μr = mean (arithmetic) return

σr = standard deviation of returns

Zα = normal distribution parameter

P t−1 = beginning period value

V AR(5%) = (−5% + 1.65 × 10%) × 200 = $23

V AR(1%) = (−5% + 2.33 × 10%) × 200 = $36.6

9
© 2014-2023 AnalystPrep.
Q.2628 You are assigned to calculate the monthly VaR for the stock of Apex Inc. You are provided
with the following data for the ten worst returns of the stock during the last 100 months:
-12%, -7%, -32%, -26%, -24%, -20%, -19%, -17%, -15%, -14%

Which of the following is closest to the monthly VaR for Apex using a confidence level of 95%?

A. -32%

B. -17%

C. -12%

D. -14.5%

T he correct answer is B.

T he 95% VaR can be found by finding the value that separates the 5% worst values of the returns
distribution from the remaining distribution. T his value will be the [(1-95%)100 + 1]th observation,
i.e., 6th observation after rearranging all the observations in ascending order.
[-32% -26% -24% -20% -19% -17% -14% -15% -12% -7%]
T hus, our observation of interest is -17%

10
© 2014-2023 AnalystPrep.
Q.2629 Jason T yler has invested $100,000 in the shares of Kraken Corp. To calculate the market risk
of his portfolio, T yler gathers the monthly returns for the security over the last 500 months. T he 10
worst returns during this period were:
-30%, -27%, -24%, -23%, -22%, -21%, -20%, -19%, -18%, -16%

What is the monthly VaR for T yler’s investment using a confidence level of 99%?

A. $30,000

B. $16,000

C. $21,000

D. $27,000

T he correct answer is C.

In general, if there are n ordered observations, and a confidence level cl%, the cl% VaR is given by

the [(1 – cl%) n + 1]th highest observation. T his is the observation that separates the tail from the

body of the distribution. For instance, if we have 1,000 observations and a confidence level of 95%,

the 95% VaR is given by the (1 – 0.95)1,000 + 1 = 51st observation. T here are 50 observations in

the tail.

In this case, the look-back window has 500 observations, so the 99% VaR is given by the (1 –

0.99)500 + 1 = 6th worst observation. T here are 5 observations in the tail.

T hus, monthly VaR = 21% * $100,000 = $21000

11
© 2014-2023 AnalystPrep.
Q.2630 Given a normally distributed profit and loss distribution with an annual mean of $500,000 and
standard deviation of $50,000, calculate the VAR at a confidence level of 99%.

A. $598,000

B. $383,500

C. $616,500

D. $402,000

T he correct answer is B.

V aR = −μ + σ × z

For a confidence level of 99%, the Z value will be 2.33.

T he VaR can be calculated as:

V aR = −500, 000 + 2.33(50, 000) = −500, 000 + 116, 500 = −383, 500

12
© 2014-2023 AnalystPrep.
Q.2632 An analyst has gathered the following information about a portfolio which has normally
distributed geometric returns:

Mean 10%
Standard Deviation 40%
Portfolio 100 million

What is the 95% lognormal VAR for this portfolio?

A. $74.7 million

B. $35.3 million

C. $42.8 million

D. $113.4 million

T he correct answer is C.

Lognormal VaR = P (1 − eμ−σz )

= 100, 000, 000(1 − e0. 1−0. 4(1. 645)
= 100,000, 000 × (−0.4276) = 42,764, 737

13
© 2014-2023 AnalystPrep.
Q.2633 Jimmy Ray, a risk analyst at Alcoa Bank, has just performed a historical simulation for
estimating the VAR for the fixed income portfolio of the bank based on the returns for the last 500
trading days. T he 10 worst one day returns generated in the simulation are:
-9,111, -8,669, -8,127, -7,098, -6,712, -6,698, -5,743, -5,189, -4,811, -4,775

Which of the following is the 99% one day expected shortfall for the portfolio?

A. 8,145

B. 6,712

C. 9,111

D. 7,943

T he correct answer is D.

From a statistical point of view, the expected shortfall, also known as the conditional VaR (CVaR), is a

sort of mean excess function, i.e. the average value of all the values exceeding a special threshold,

the VaR. CVAR indicates the potential loss if the portfolio is “hit” beyond VAR:

ES = E[L|L is greater than VaR]

If there are n ordered observations, and a confidence level cl%, the cl% VaR is given by the [(1 –

cl%) n + 1]th highest observation. In this case, the 99% VaR is given by the (1 – 0.99)500 + 1 = 6th

worst observation. T here are 5 observations in the tail.

T he 99% expected shortfall will be the average of the 5 worst returns (tail losses) which in this case

will be:

(−9, 111 + −8, 669 + −8, 127 + −7, 098 + −6, 712)
= 7, 943.4
5

14
© 2014-2023 AnalystPrep.
Q.2636 Jacob Watson is a risk manager for a large bank. Presently, he is estimating the VaR for the
equities portfolio of the bank. He is considering estimating the VaR using both a normal distribution
assumption and a lognormal distribution assumption. He has gathered the following information about
the portfolio:

Value of the portfolio USD 1 million

Mean 15%
Volatility 25%

What would be the 1 year 99% VaR for the portfolio under the two assumptions?

A. Normal distribution: $495,000; Lognormal distribution: $460,000

B. Normal distribution: $460,000; Lognormal distribution: $495,000

C. Normal distribution: $432,500; Lognormal distribution: $351,000

D. Normal distribution: $499,000; Lognormal distribution: $432,500

T he correct answer is C.

Under normal distribution assumption:

V aR = μ − σ × z

For a confidence level of 99%, the Z value will be 2.33.

T he VaR can be calculated as:

V aR = 0.15 − 2.33(0.25) = 43.25%

In dollar terms, this will be $432,500.

Under the lognormal distribution assumption:

Lognormal VaR = 1 − eμ−σ×z = 1– e0. 15–0. 25(2. 33) = 35.1%

In dollar terms, this will be $351,000.

15
© 2014-2023 AnalystPrep.
Q.2815 If the geometric return R t over some period of time is 8.62%, what is the arithmetic return?

A. 0.0831

B. 0.0862

C. 0.0255

D. 0.0900

T he correct answer is D.

Recall that,

R t = ln(1 + rt )
⇒ rt = eRt − 1 = e0. 0862 − 1 = 0.09002 ≈ 9%

Q.2816 If the geometric returns R t is 0.013 over a specified period, then the arithmetic returns
must be:

A. 0.032562

B. 0.245861

C. 0.013085

D. 0.056894

T he correct answer is C.

Geometric Return = ln[1 + Arithmetic Return]

Let rt be the arithmetic return

0.013 = ln(1 + rt )
exp0. 013 = 1 + rt
rt = exp0. 013 − 1 = 0.013085

16
© 2014-2023 AnalystPrep.
Q.2817 Assuming that the P/L over a specified period is normally distributed and has a mean of 14.1
and a standard deviation of 28.2. What is the 95% VaR and the corresponding 99% VaR?

A. T he 95% VaR is 32.289 and the 99% VaR is 51.4932

B. T he 95% VaR is 36.495 and the 99% VaR is 51.556

C. T he 95% VaR is 55.236 and the 99% VaR is 36.49551

D. T he 95% VaR is 36.225 and the 99% VaR is 41.586

T he correct answer is A.

Recall that:

αVaR = −μP /L + σP/L zα

T herefore, the 95% VaR is: −14.1 + 28.2Z0. 95 = −14.1 + 28.2 × 1.645 = 32.289

T he 99% VaR is: −14.1 + 28.2Z0. 99 = −14.1 + 28.2 × 2.326 = 51.4932

17
© 2014-2023 AnalystPrep.
Q.2818 T he arithmetic returns rt , over some period of time, are normally distributed with a mean of
1.55 and a standard deviation 1.07. T he portfolio is currently worth 1 unit. Calculate the 95% VaR and
the 99% VaR.

A. T he 95% VaR is 2.3658 and the 99% VaR is 3.6588

B. T he 95% VaR is 1.4542 and the 99% VaR is 0.0652

C. T he 95% VaR is 0.6742 and the 99% VaR is 3.00896

D. T he 95% VaR is 0.21015 and the 99% VaR is 0.93882

T he correct answer is D.

Recall,

αVaR = −(μr − ðrzα )P t−1

T herefore, the 95% VaR is: −1.55 + 1.07 × 1.645 = 0.21015

T he 99% VaR is: −1.55 + 1.07 × 2.326 = 0.93882

18
© 2014-2023 AnalystPrep.
Q.2819 Let’s assume that the geometric returns R t , are normally distributed with mean 0.079 and
standard deviation 0.312. Further, assume that the portfolio is currently worth 1 unit. Calculate the
95% and 99% lognormal VaR.

A. T he 95% VaR is 0.8951 and the 99% VaR is 0.2351

B. T he 95% VaR is 0.88526 and the 99% VaR is 0.56898

C. T he 95% VaR is 0.3522 and the 99% VaR is 0.4762

D. T he 95% VaR is 0.8951 and the 99% VaR is 0.56898

T he correct answer is C.

From the lognormal derivation,

αV aR = P t−1 − P ∗ = P t−1(1 − exp[μR − σR Zα ])

Applying the formula in the question we have:

95% V aR = 1 − exp(0.079 − 0.312 × 1.645) = 0.3522

99% V aR = 1 − exp(0.079 − 0.312 × 2.326) = 0.4762

Q.3006 If the arithmetic returns rt over some period of time are 0.09, what are the geometric
returns?

A. 0.0831

B. 0.08618

C. 0.02546

D. None of the above

T he correct answer is B.

Recall that, R t = ln(1 + rt )

T herefore, ln(1 + 0.09) = ln(1.09) = 0.08618

19
© 2014-2023 AnalystPrep.
Q.3011 Assume you are dealing with a stock “A” that displays a highly negatively skewed distribution
comprised of the past 260-days returns. Suppose you have P1 = A and P2 = -A, meaning P1 is long
stock A and P2 is short stock A. Which statement is most likely to be accurate about a 99% VaR?

A. VaR (P1) > VaR (P2)

B. VaR (P1) < VaR (P2)

C. VaR (P1) = VaR (P2)

D. Cannot be concluded from the given information

T he correct answer is A.

Given that the return distribution of stock A is negatively skewed, the implication is that it displays a

long left tail, meaning large potential losses for a long position and conversely large potential gains for

a short position. T herefore, VaR (P1) will be expected to be higher. If the distribution was

symmetric, both P1 and P2 would have the same VaR.

Addi ti onal Expl anati on

A long left tail means there's a high probability of large potentially crashing losses on the asset. If you

are long the asset, then it means you are exposed to more risk; the asset price might fall

considerably, triggering a major loss. If you sell short (borrow and sell in the hope of repurchasing

the asset at a later date) your repurchase price is likely to be lower, which means you will record a

gain. Similarly, a short position in a call option on the asset will most likely make a gain (keep the

premium) because the asset will most likely fall in price, and the the long position won't exercise

their right to buy.

20
© 2014-2023 AnalystPrep.
Q.3036 What would be the 95% parametric VaR of a portfolio made of two independently normally
distributed stocks – A and B, with A ∼ N(0.5, 1) and B ∼ N(3, 15). Assume that P = (A + B)

A. 56 × P

B. 4.87 × P

C. 3.08 × P

D. None of the above

T he correct answer is C.

Assuming (A + B) , then P ∼ N (3.5, 16).

V aR(α%) = [−μr + σr × Zα ] P = [−3.5 + 4 × 1.645] P =3.08P

Note: T he sum of two independent normally distributed random variables is normal, with its mean

being the sum of the two means, and its variance being the sum of the two variances

Where σr = √16 = 4

21
© 2014-2023 AnalystPrep.
Q.5037 During a risk management meeting, a stock analyst presents a report. T he report includes a
Q-Q plot comparing actual return quantiles with normal distribution quantiles. According to the
analyst, the team's assumption of normal distribution in their previous analyses is erroneous because
the data shows characteristics of a Student's t-distribution. In regard to the Q-Q plot, which of the
following evidence supports the analyst's claim?

A. T he Q-Q plot has a negative y-intercept

B. T he Q-Q plot is linear

C. T he tails and central mass observations show steeper slopes

D. T he plot's slope is negative

T he correct answer is C.

Steeper slopes on the tails and central observations that appear to be skinnier than that of the

reference (normal) distribution would all indicate that the empirical distribution is non-normal. T he

fat tails mean that extreme events occur more frequently in reality than what a normal distribution

would predict.

More on QQ Pl ots
Q-Q plots, or quantile-quantile plots, are statistical tools that help us determine whether or not our
data is likely to have come from some theoretical distribution, such as the Normal or exponential. As
an example, we can confirm if our dependent variable is normally distributed by using a Normal Q-Q
plot. Despite the plot providing a quick visual check, it is not a perfect test, so it is quite subjective.
However, it helps us to quickly establish whether our assumptions are plausible, and if not, how the
assumption was violated and which data points contributed to the violation.

22
© 2014-2023 AnalystPrep.
 Reading 64: Non-parametric Approaches

Q.1484 In order to compile historical simulation data, one would need to assemble historical P/L or
return observations on the positions in the current portfolio. T he series of historically simulated P/L
would form the basis of historically simulated VaR and ES. Which of the following would be correct
regarding the simulated series?
I. T his series would not be same as the actual P/L earned on the portfolio.
II. It would show the P/L earned on the current portfolio.
III. T his series would give an accurate measurement of the actual P/L.

A. Point I

B. Point II

C. Point III

D. None of the above

T he correct answer is A.

T here would be a difference in actual P/L and historically simulated P/L because the portfolio
changes over time and is subjected to mapping approximations.

Q.1485 T here are different benefits of parametric methods over nonparametric methods. An FRM
manager will consider non-parametric methods as the most natural choice for high-dimension
problems. Which of the following is NOT an advantage of nonparametric methods?

A. It is easy for non-parametric methods to accommodate high dimensions.

B. In non-parametric methods, there is no difficulty in dealing with variance-covariance

matrices and curses of dimensionality.

C. T he non-parametric approaches can accommodate skewness, fat tails and other non-
normal features.

D. T he results from non-parametric methods are often hard to communicate.

T he correct answer is D.

It is very easy for non-parametric methods to accommodate high dimensions; there are no problems
in dealing with variance-covariance matrices; and they also use data that is readily available. In
addition, results are easy to communicate.

23
© 2014-2023 AnalystPrep.
Q.1486 T here are two bootstrapping approaches: parametric and non-parametric.
I. For a “non-parametric bootstrap”, the exercise proceeds from a given data sample that’s
representative of the population. By re-sampling the data, you can reproduce the distribution of your
estimator, or just its mean, variance, etc.
II. In a “parametric bootstrap” you have an assumption on the underlying distribution of the
population up to some parameter. What you do is then estimate the parameter from the data, and
then sample from the assumed distribution with the estimated parameter.
III. In a “non-parametric bootstrap” we make assumptions about how the sample size will be
distributed and resample the parameter.

Which of the following statements correctly define each of these approaches?

A. Point I and III

B. Point I and II

C. Point II and III

D. None of the above

T he correct answer is B.

Nonparametric bootstraps make no assumptions about how your observations are distributed and

actually involve “re-sampling” your original sample. Parametric bootstraps resample a known

distribution function, whose parameters are estimated from your sample.

Further Explanation

In a nonparametric bootstrap, a sample of the same size as the data is take from the data with

replacement. Let's say you have an original sample comprised of 30 observations. By sampling with

replacement, you can create a new sample of size 30 by replicating some of the observations in the

original sample and omitting others.

Parametric bootstrapping, on the other hand assumes that the data comes from a known distribution

with unknown parameters. For example the data may come from a normal, Poisson, negative

binomial for counts, or some other continuous distribution. You then estimate the parameters from

the data that you have and then you use the estimated distributions to simulate the samples.

24
© 2014-2023 AnalystPrep.
Q.1487 T he non-parametric density estimation is based on the assumption that a basic historical
simulation does not get the best out of the information at hand. Which of the following examples
demonstrates this drawback?

A. If we have 100 P/L observations, the basic HS only permits us to estimate VaR at discrete
confidence levels, say, 95%.

B. If we have 100 P/L observations, the VaR at the 95% confidence level is given by the
seventh-largest loss.

C. If we have 100 P/L observations, the VaR at the 95% confidence level is given by the
fourth-largest loss.

D. None of the above.

T he correct answer is A.

Historical simulation does not allow us to estimate the VaR at non-discrete intervals. If we have 100
P/L observations, the VaR at 95% confidence level is given by the sixth-largest loss, but the VaRs at
95.1%, 95.9%, and 95.5% confidence level cannot be obtained because there are no corresponding
loss observations.

Q.1488 Which of the following methods is NOT used to represent data under non-parametric density
estimation?

A. Histogram

B. Bar Chart

C. Naive Estimators

D. Kernels

T he correct answer is B.

Bar charts are not used for showcasing data as they do not justify the observations adequately.

25
© 2014-2023 AnalystPrep.
Q.1489 All of the following items are generally considered advantages of non-parametric estimation
methods except:

A. little or absolute lack of reliance on covariance matrices.

B. Use of historical data.

C. Ability to accommodate largely skewed data.

D. Availability of data.

T he correct answer is B.

Use of historical data is actually a disadvantage of non-parametric estimation methods because the
past is not necessarily an indication of the future. In particular, data gathered from a relatively quiet
(volatile) past period may lead to the development of models that underestimate (overestimate)the
current risk level.

26
© 2014-2023 AnalystPrep.
Q.1490 Estimating historical simulation ES or VaR does not have any theoretical problems; however,
it has a practical problem. Which one is it?

A. As the holding period decreases, the number of observations decreases too.

B. As the holding period increases, the number of observations decreases.

C. As the holding period decreases, the size of data decreases.

D. As the holding period increases, the size of the data decreases.

T he correct answer is B.

As the holding period rises, the number of observations rapidly falls, and we soon find that we don't

have enough data. Let's illustrate:

If we have 1000 observations of daily P/L, that corresponds to four years' worth of data at 250

trading days a year. What does this mean?

I. We have 1000 P/L observations if we use a daily holding period.

II. We have 200 observations if we use a weekly holding period with 5 days to a week because
each weekly P/L will be the sum of five daily P/Ls
III. We have only 50 observations if we use a monthly holding period, again because each
monthly P/L will be the sum of 20 daily P/Ls
... and so on....and so on....

As such, it is clear that the number of effective observations rapidly falls as the holding period rises,

and the size of the data set imposes a major constraint on how large the holding period can practically

be.

C states that as the holding period decreases, the size of data decreases. T hat's correct, but it is

hardly a problem. T hat would actually be an advantage because we would have more data points that

are easier to gather.

27
© 2014-2023 AnalystPrep.
Q.1491 T he theory of order statistics gives us a complete function of VaR (or ES) distribution and it’s
a very easy-to-compute-and-apply approach. Which of the following statements is INCORRECT
regarding estimates of the standard normal VaR based on the conventional formula?

A. As n increases, the estimated VaR median tends to converge to conventional estimate.

B. T he confidence interval is wide as n gets larger and narrow for low values of n.

C. T he confidence interval shows a gap between 5 and 95 percentiles.

D. T he confidence interval is wider for extreme VaR confidence levels.

T he correct answer is B.

Recall that that with n observations, we take the VaR as equal to the negative of the rth lowest
observation,
where

r = n(1 − α) + 1

Also, note that estimates of the lower and upper percentiles of the VaR distribution function give the
estimates of the bounds of our VaR confidence interval. T hus, the confidence interval is quite wide
for low values of n, but narrows as n gets larger.

28
© 2014-2023 AnalystPrep.
Q.1492 Which on of the following statements is most likely correct? A bootstrapping exercise:

A. resampling from our existing data set without replacement.

B. assumes that the distribution of returns will remain the same in the past and in the future.

C. assumes that the distribution of returns in future will be markedly different from past
distributions.

D. results in a VaR estimate that is a sum of sample VaRs after repeated sampling.

T he correct answer is B.

Bootstrapping presents a simple but powerful improvement over basic Historical Simulation is to

estimate VaR and ES. Crucially, it assumes that the distribution of returns will remai n the same in

the past and in the future, justifying the use of historical returns to forecast the VaR.

A bootstrap procedure involves resampling from our existing data set with replacement. A sample is

drawn from the data set, its VaR recorded, and the data “returned.” T his procedure is repeated over

and over. T he final VaR estimate from the full data set is taken to be the average of all sample VaRs.

In fact, bootstrapped VaR estimates are often more accurate than a ‘raw’ sample estimates.

Q.1493 T he closed-form confidence intervals have limited applicability as we do not have

expressions for standard errors. Which of the following parameters will pose similar problems?

A. Medians

B. Tail probabilities

C. Correlations

D. All of the above

T he correct answer is D.

To estimate confidence intervals for all the parameters above, traditional closed-form approaches
would require us to make the use of statistical theory, but the theory itself is of limited use.

29
© 2014-2023 AnalystPrep.
Q.1494 In layman’s term, there is a huge problem with volatility as a measurement of risk, at least
for the returns of financial securities. T he problem is that excess returns are seen as risky. T his is
why in the 80s, J.P. Morgan came with a measurement of risk that only considered downside risk, the
Value at Risk (VaR). One of the methods used to measure the VaR is bootstrapping. Which of the
below statements is an advantage of bootstrapping?
I. Bootstrapping allows us to estimate confidence intervals of any parameter.
II. T he underlying assumptions behind bootstrapping are less demanding.

A. Both Point I and II

B. None of the above

C. Point I

D. Point II

T he correct answer is A.

Both I and II are advantages of bootstrapping. Bootstrapping allows us to estimate confidence

intervals of any parameter and the underlying assumptions behind bootstrapping are less demanding.

Q.1495 Even though bootstrapping has numerous advantages, the bootstrap estimates are associated
with a little bias or error. Which of the following presents an error of bootstrapping?

A. Un-sampling variability.

B. Re-sampling variability.

C. Dual-sampling variability.

D. Bootstrapping variability.

T he correct answer is B.

T here are two types of errors in bootstrapping estimates: sampling variability (we have a sample of
size n drawn from our population, rather than the whole population itself); and resampling variability
(we take only B bootstrap re-samples rather than an infinite number)

30
© 2014-2023 AnalystPrep.
Q.1496 One of the drawbacks of the historical simulation approach is that the discreteness of the
data rules out estimation of VaRs between data points. For example, if there were 100 historical
observations, estimation of the VaR is a straightforward process at the 95% or the 96% confidence
levels, but it is impossible to incorporate a confidence level of, say 95.5%. Which of the following
methods can solve this problem?

A. Applying Brute Force

B. Bootstrapping

C. non-parametric density estimation

D. Use of a large number of re-samples

T he correct answer is C.

Simple HS only allows VaR estimate at discrete confidence levels. Non-parametric density estimation
treats data as if they were drawings from some unspecified or unknown empirical distribution
function. T he existing data points can be used to “smooth” the data points, paving the way for VaR
calculation at all confidence levels.

Q.1497 Which of the following is NOT a step from the three-step model developed by Andrews and
Buchinsky to determine B (the number of bootstrap resamples)?

A. Assuming a value “x” and plugging it into the relevant equation to get a preliminary value of
B.

B. Simulating B0 re-samples.

C. Taking an infinite number of bootstrap re-samples.

D. Making an estimate of the sample kurtosis.

T he correct answer is C.

We cannot take an infinite number of re-samples; only a finite (desired) number is used.

31
© 2014-2023 AnalystPrep.
Q.2631 An analyst performing a historical simulation to measure the VaR of a portfolio uses a
volatility-weighted approach. One month ago, the actual return of the asset was 5% and its daily
volatility estimate was 2%. If the current daily volatility is 1.5%, calculate the volatility-adjusted
return.

A. 0.03

B. 0.0167

C. 0.0375

D. 0.0667

T he correct answer is C.

σT ,i 0.015
r∗t,i = ( ) rt,i = ( ) 0.05 = 0.0375
σt,i 0.020

Q.2635 Which of these statements about non-parametric approaches for estimating VAR is not true?

A. Non-recurrent extreme losses can dominate and skew non-parametric estimates.

B. If there is a lot of volatility in the period chosen, the VAR and expected shortfall estimates
will be too high.

C. Unlike parametric approaches, non-parametric approaches can easily accommodate non-

normal features like fat tails and skewness.

D. Data for these methods can be very difficult to obtain and needs frequent adjustments.

T he correct answer is D.

Non-parametric approaches are easy and intuitive to use and do not require any assumptions for
probability distributions or variance-covariance matrices. T hey rely primarily on historical data to
estimate VAR. Data for these approaches is easy to obtain and does not require adjustment. Since
these approaches rely on historical data only, estimates can be skewed by periods of high volatility or
by non-recurrent extreme losses.

32
© 2014-2023 AnalystPrep.
Q.2820 T he mean return from a dataset has been pre-calculated and is given as 0.04. T he standard
deviation has also been given as 0.32. With 90% confidence, what will be our percentage maximum
loss? Assume that from our dataset, Z = -1.28 and N(Z) = 0.10 since you are to locate the value at
the 10th percentile.

A. 36.96%

B. 11.27%

C. 11.32%

D. 36.72%

T he correct answer is A.

Recall that

(X − μ)
Z=
σ

From the data, we are given that: μ = 0.04, σ = 0.32, and Z = −1.28

(X−0. 04)
T herefore: −1.28 = ⇒ X = −1.28(0.32) + 0.04 = −0.3696
(0. 32)

X = −0.3696 = 36.96% loss

T his means that we are 90% confident that the maximum loss will not exceed 36.96%.

33
© 2014-2023 AnalystPrep.
Q.2821 Which of the following is NOT an advantage of non-parametric methods?

A. When combined with add-ons, non-parametric methods are capable of refinement and
potential improvement.

B. Non-parametric methods can accommodate any type of position including derivative

positions.

C. Non-parametric methods are not subject to the phenomenon of ghost effect or shadow
effects.

D. Non-parametric methods can accommodate fat tails, skewness and other abnormal
features to parametric approaches.

T he correct answer is C.

Most (if not all) non-parametric methods are subject (to a greater or lesser extent) to the
phenomenon of ghost or shadow effects. T his is actually a disadvantage.

Q.2822 T here are 30 observations in a dataset. T he worst 10 return observations (in %) are listed
below:
[-20, -18, -18, -17, -15, -14, -12, -10, -7, -3]

What is the VaR at the 90% confidence?

A. 17%

B. 18%

C. 16%

D. 15%

T he correct answer is A.

T he VaR at 90% cl is given by the [(1 − 0.90)30 + 1] = 4th worst observation

V aR = 17%

34
© 2014-2023 AnalystPrep.
Q.2831 Find the weight of an observation 13 days ago if the total number of days in the historical
window is 300 with a 0.8 control rate of memory decay.

A. 0.013

B. 2.2050

C. 0.0205

D. 0.2250

T he correct answer is A.

T he weight of observation i-days ago is given by:

λi−1(1 − λ)
w(i) =
1 − λn

Where n is the number of days in the historical window and θ is the control rate of the memory

decay,

T herefore:

0.812 (1 − 0.8)
W (12) = = 0.01374
1 − 0.8300

35
© 2014-2023 AnalystPrep.
Q.3007 You are a VaR analyst on a trading floor and one of the traders has complained to you about
the historical VaR of one of his stock: he argues that over the past 3 trading days, the end–of-trading
stock price has been flat and as a result, no movement is expected from his VaR. You noted however
that though the stock price has remained flat, interest rates have moved. In these circumstances, is
the trader right?

A. Yes, the trader is right: as long as the stock price didn't move, the risk remains unchanged.

B. No, the trader is wrong: even if the stock price did not move, then the risk of that stock
should change.

C. Yes, the trader is right: price is the only determinant of daily VaR.

D. None of the above: further analysis is required.

T he correct answer is D.

Further analysis is required: the historical VaR is based on a set of moving 260 days return. As a
result one of your previous day worst scenarios might move out of your 260 days window. T he
result of this will be that though your stock price may remain flat, your tail of worst return might
change, resulting in a VaR variation.

Q.3010 You have been hired on a popular trading floor and one of the traders comes over and asks
about the impact of price changes on her VaR - made of a single long position in stock KKJL.
Yesterday's closing price was USD 100.
You are using a 95% confidence historical VaR based on a 260 days moving window of historical data.
In this time period, the 16 worst 1-day returns from for KKJL as of yesterday were as follows (in %):
-9.5, -8, -7.6, -7.4, - 7.2, -7.18, -7.1, -6.9, -6.57, -6.56, -6.45, -6.4, -6.25, -6.05, -5.99, -5.85.

Suppose that the stock price decreased by 10% between yesterday and today following the
publication of an adverse dossier on the company. T he latest return to slip out of the 260-day moving
window is -3%.

What will be the 95% confidence VaR in absolute terms?

A. USD 6.25

B. USD 5.625

C. USD 6.625

D. Cannot be concluded from the given information

T he correct answer is B.

36
© 2014-2023 AnalystPrep.
In general, if there are n ordered observations, and a confidence level cl% , the cl% VaR is given by

the [(1 – cl%) n + 1]th highest observation. T his is the observation that separates the tail from the

body of the distribution. Given 260 observations, we are therefore interested in the [(1 - 0.95)260 +

1]th = 14th worst observation

Before the latest 10% loss, the ordered losses are as follows:

[-9.5, -8, -7.6, -7.4, - 7.2, -7.18, -7.1, -6.9, -6.57, -6.56, -6.45, -6.4, -6.25, -6.05, -5.99, -5.85]

T he 14th worst observation is -6.05. However, given that the stock decreased by 10% between

yesterday and today, the arrangement changes; -10% will now be the worst return that the stock

experienced over the last 260 business days. T he 16 worst returns shall now be:[-10, -9.5, -8, -7.6,

-7.4, - 7.2, -7.18, -7.1, -6.9, -6.57, -6.56, -6.45, -6.4, -6.25, -6.05, -5.99]

T he VaR at 95% will be based on -6.25%, which is now the 14th worst return.

V aR = 90 × 6.25% = 5.625

Notes:

(I) 90 here is the price of the stock today (after a 10% decline yesterday)
(II) We're working with the worst returns recorded over a 260 days moving window, so we're
essentially looking back and ordering the worst returns recorded during that period, and one
observation slips out of the window every day. T he latest return to slip out is -3%, but this is
considerably higher (less negative) than -5.85% (lowest worst return on our list), so it does not
affect the makeup of the worst 16 returns.

37
© 2014-2023 AnalystPrep.
Q.3012 Under age-weighted historical simulation,

A. more recent observations are weighted more and distant observations weighted less.

B. all observations are weighted equally.

C. the decay parameter takes a value of 1.

D. the historical window of observations must not exceed 250 days.

T he correct answer is A.

T he age-weighted historical simulation method makes an adjustment to the equal-weighted

assumption used in historical simulation; More weight is attached to recent observations. Distant
observations that are less likely to recur have less weight.
T he decay parameter takes a value between 0 and 1, with values close to 1 indicating slow decay
T he size of the historical window is not limited.

38
© 2014-2023 AnalystPrep.
Q.3035 Paul is using the age-weighted historical simulation approach to estimate the VaR of a stock
portfolio, with a historical sample size of 100 days and a decay factor of 0.96. Over the recent past,
the portfolio has registered the following returns:

Return Periods Ago(Days)

−3.2% 109
−3.3% 75
−2.3% 66
−1.3% 22
−3.0% 45

Determine the weight on the return earned 45 days ago

A. 0.05

B. 0.0025

C. 0.0065

D. 0.006751

T he correct answer is D.

Under age-weighted (aka Hybrid) historical simulation, the weight, w i, given to an observation i days

old is given by:

λi−1 (1 − λ)
wi =
(1 − λn )

So,

0.9645−1(1 − 0.96)
w 45 = = 0.006751
(1 − 0.96100)

Note: Any return falling outside the historical window would have a weight of zero, for instance, the
observation made 109 days ago.

39
© 2014-2023 AnalystPrep.
Q.3037 You have been hired on the trading floor and one of the traders comes over and asks about
the impact of a price change on her VaR made of a long position in stock A, whose value stood at 100
as of yesterday.
Assume you are using a 95% confidence historical VaR (based on 260 days moving window of
historical data). Further, assume that the 16 worst 1-day returns of stock as of yesterday were as
follows:

-9.5, -8, -7.6, -7.4, - 7.2, -7.18, -7.1, -6.9, -6.57, -6.56, -6.45, -6.4, -6.25 -6.05, -5.99,-5.85.

Assume the price of the stock increased by 10% between yesterday and today. Further, assume that
the oldest return is not among the returns given. What will the value of today's 95% VaR (in absolute
value)?

A. $6.25

B. $6.655

C. $10

D. Cannot conclude

T he correct answer is B.

Today's stock price is $100 × (1 + 10%) = $110

T he 95% VaR is given by the 14th worst return, i.e., -6.05%,

N/B: Using 260 days moving window of historical data, the 95% VaR will be r=260(1-95%) +1 = 14th
observation

T he new 95% VaR will be $110 - $110 × (1 - 6.05/100) = -$6.655

Note: T he latest return (10% ) does not effect the left tail of the loss distribution. It is higher (more

positive) than all of the returns given and does not get a spot among the worst 16 observations. In

addition, the examiner assumes that the oldest return pushed out the rolling window is not among the

entires given. T herefore, the today's worst 16 observations will be the same as yesterday's.

40
© 2014-2023 AnalystPrep.
Q.3039 T wo newly recruited risk analysts at Capital Investment Bank, Netherlands, are comparing
the calculation of the VaR under parametric and non-parametric approaches. However, they can't
quite agree on the method to use because of some of the characteristics of the loss distribution data.
All of the following characteristics would make non-parametric approaches the favored method,
EXCEPT :

A. Fat tails at both ends of the distribution

B. Skewness in the distribution

C. Heteroskedasticity in the data

D. Scarce high-magnitude loss events

T he correct answer is D.

Non-parametric approaches to VaR estimation do not make restrictive assumptions about the

underlying distribution. Parametric methods, on the other hand, assume very specific forms such as

normal or lognormal distributions.

Non-parametric approaches can therefore accommodate fat tails, skewness, and any other salient

non-normal features that could be present. However, if the data collected belongs to a period with

few losses or losses with low magnitude, non-parametric methods tend to produce risk measures that

are too low. In other words, quiet data periods will lead to VaR (as well as ES) estimates that are too

low. Hence, parametric approaches would be more appropriate in such a situation.

41
© 2014-2023 AnalystPrep.
 Reading 65: Parametric Approaches (II): Extreme Value

Q.2175 Extreme events have a very low probability of occurrence but are nonetheless taken very
seriously in the financial world. Which of the following best explains why?

A. Extreme events tend to recur at rather regular time intervals.

B. Extreme events rarely have warning signs and thus markets cannot prepare for them in
any way.

C. Extreme events are normally very costly and can create a “ripple” effect on the global
market.

D. No models have been developed to accurately predict and estimate the effects of extreme
events in qualitative terms.

T he correct answer is C.

Although extreme events have extremely low associated probabilities, they are normally high-
impact, heavy-loss events. T hey rarely occur, but when they do, their impact is so dramatic and can
lead to failure of key market players and loss in value of key market products with a huge
subscription base. In some cases, extreme events can trigger off a financial crisis.

Q.2176 It’s normally very difficult and problematic to model extreme events mai nl y because of:

A. A lack of models that can estimate the effects of certain extreme but possible events.

B. A lack of qualified personnel to oversee the modeling process.

C. A lack of credible, reliable input data.

D. T he fact that extreme event modeling requires a considerable investment of time and
expertise.

T he correct answer is C.

T here are very few historical observations on which to base our estimates. A model is only as good
as the input data. Due to a lack of credible historical data on extreme events (some of which have
never occurred but can still occur), practitioners formulate assumptions. Unfortunately, some of the
assumptions are normally out of sync with the reality, meaning the resulting estimates are also not
reliable. For example, an incorrect assumption regarding the distribution of a certain phenomenon
might correctly model central observations but fail to come up with reliable estimates of extreme
observations.

42
© 2014-2023 AnalystPrep.
Q.2177 Extreme events can best be modeled through the application of:

A. T he central limit theorem

B. T he standard normal distribution

C. Extreme-value theorems

D. T he exponential distribution

T he correct answer is C.

Central tendency statistics are best governed by central limit theorems, but extreme tendency
statistics can only be governed, appropriately enough, by extreme-value theorems such as the widely
applied Fisher-T ippett theorem.

Q.2178 T he following is the probability distribution function of the generalized extreme value
distribution:

1
⎧
⎪ −
⎪ exp [−(1 + ξ(X−μ) ) ξ ] ,
⎪ ξ ≠0
σ
H ξ,μ ,α =⎨
⎪
⎩ exp [−exp (− (X−μ) )] ,
⎪
⎪ ξ =0
σ

ξ(X−μ)
Where X satisfies the condition 1 + > 0 What is represented by μ,α, and ξ respectively?
σ

A. T he location parameter, the scale parameter, and the shape parameter

B. T he scale parameter, the location parameter, and the shape parameter

C. T he mean, the variance, and the shape parameter

D. T he mean, the shape parameter, and the location parameter

T he correct answer is A.

T he GEV distribution has 3 parameters: μ represents the location parameter of the limiting
distribution, α represents the scale parameter, and ξ represents the shape parameter (it gives an
indication of the heaviness of the tail of the limiting distribution).

43
© 2014-2023 AnalystPrep.
Q.2179 T he following is the probability distribution function of the generalized extreme value
distribution:

1
⎧
⎪ −
⎪ exp [−(1 + ξ(X−μ) ) ξ ] ,
⎪ ξ ≠0
σ
H ξ,μ ,α =⎨
⎪
⎪
⎩ exp [−exp (−
⎪ (X−μ)
)] , ξ =0
σ

ξ(X−μ)
Where X satisfies the condition 1 + >0
σ
If ξ > 0, the GEV becomes the:

A. Frechet distribution

B. Pareto distribution

C. Gumbel distribution

D. Weibull distribution

T he correct answer is A.

If ξ > 0, the GEV becomes the Frechet distribution and applies when the cumulative distribution of
X obeys a power function as is therefore heavy.

Q.2180 If ξ < 0, the GEV becomes the Weibull distribution, but this distribution is rarely used to
model financial returns mainly because:

A. Its cumulative distribution has heavier than normal tails and very few empirical financial
returns are heavy-tailed.

B. Its cumulative distribution has lighter than normal tails and very few empirical financial
returns are light-tailed.

C. It’s asymmetric.

D. It’s symmetric.

T he correct answer is B.

If ξ < 0, the GEV becomes the Weibull distribution, corresponding to the case where F(x) has lighter
than normal tails. T his is precisely why the Weibull distribution is not used to model most empirical
financial returns since only a few of them have light tails.

44
© 2014-2023 AnalystPrep.
Q.2182 For the standardized Frechet distribution with ξ=0.3, the 5% quantile is equal to:

A. -0.9000

B. -0.8340

C. -0.4567

D. -0.9349

T he correct answer is D.

When ξ > 0, X = μ − ( σξ ) [1 − (−ln(p))−ξ ]

T herefore, the 5% quantile = −( 0.13)[1 − (−ln(0.05))−0. 3] = −0.9349

45
© 2014-2023 AnalystPrep.
Q.2956 Suppose that we are informed that in a U.S. stock market, the location parameter of the
limiting distribution, μ, is 2%, the scale parameter, σ, is 0.6, and the tail index, ξ , is 0.4. Apply these
parameters in the Fréchet VaR formula to calculate the estimated 95% VaR and 99.5% VaR,
respectively. Assume n = 100.

A. 95% VaR: 1.340; 99.5% VaR: 1.657

B. 95% VaR: 1.657; 99.5% VaR: 1.119

C. 95% VaR: 1.28; 99.5% VaR: 2.477

D. 95% VaR: 1.657; 99.5% VaR: 1.876

T he correct answer is C.

Recall that, for ξ > 0:

σn
V aR = μn − [1 − (−nln(α))−n ξ ]
ξn

T herefore,

For 95% VaR:

0.6
V aR = 2 − [1 − (−100 × ln(0.95))−0. 4] = 1.28
0.4

For 99.5% VaR:

0.6
V aR = 2 − [1 − (−100 × ln(0.995))−0. 4] = 2.477
0.4

46
© 2014-2023 AnalystPrep.
Q.2957 Assuming that we are given the following parameters based on the empirical values from
contracts on futures clearing companies; β = 0.7, xi = 0.12, u = 3% , N u /n = 5% . Compute the VaR
and the Expected Shortfall at 99.5%, respectively.

A. VaR: 1.674; Expected Shortfall: 2.453

B. VaR: 4.856; Expected Shortfall: 5.905

C. VaR: 1.453; Expected Shortfall: 2.420

D. VaR: 1.667; Expected Shortfall: 2.554

T he correct answer is B.

Recall that:

−ξ
β n
V aR = u + {[ (1 − α)[] − 1}
ξ Nu

And

V aR β − ξu
ES = +
1− ξ 1 −ξ

T herefore:

−0. 12
0.7 1
V aR = 3 + {[ (1 − 0.995)[] − 1} = 4.856
0.12 0.05

4.856 0.7 − 0.12 × 3

ES = + = 5.905
1 − 0.12 1 − 0.12

47
© 2014-2023 AnalystPrep.
Q.3993 To retrieve the value at risk (VaR) for the U.S stock market under the generalized extreme-
value (GEV) distribution, a risk analyst uses the following equation which characterizes a heavy-
tailed Fréchet distribution.

σn
VaR α = μn − [1 − (−nln (α))−ξn ]
ξn

T he analyst uses the following somewhat "realistic" parameters:

Location, μ = 3.0%

Scale, σ = 0.70

Tail index, ξ = 0.30

If the sample size, n = 100, then which is nearest to the implied 99.90% VaR?

A. 2.3651%

B. 2.547%

C. 3.521%

D. 5.3216%

T he correct answer is D.

σn
VaR α = μn − [1 − (−nln (α))−ξn ]
ξn

At α = 0.999,

0.7
VaR 0. 999 = 3 − [1 − (−100ln(0.999))−0. 3 ] = 5.3216%
0.3

48
© 2014-2023 AnalystPrep.
Q.3994 In FRM parlance, an extreme value is one that has a low probability of occurrence but
potentially disastrous (catastrophic) effects. T he main challenge posed by extreme values is that:

A. T hey do not conform to any of the established loss distributions

B. T hey can only be fully characterized by multiple loss distributions

C. T hey are too big such that the resulting loss estimates are infinitely large

D. T here are only a few observations from which a credible, reliable analytical model can be
built

T he correct answer is D.

T he main challenge posed by extreme values is that there are only a few observations from which a
credible, reliable analytical model can be built. In fact, there are some extreme values that have
never occurred but that does not necessarily imply there’s no chance of occurrence in the future.
T rying to model such events can be quite an uphill task.

Q.3995 Simon is trying to fit a distribution to the extreme loss tail of a historical financial return
dataset. He does not have a good fit for the parent distribution, and has therefore not settled on an
appropriate extreme value theory (EVT ) approach. He decides to conduct a hypothesis test and
concludes that the tail index is insignificant. In this scenario, which of the following distributions
should he use?

A. Gumbel

B. Fréchet

C. Normal

D. None of the above

T he correct answer is A.

One practical consideration the risk managers and researchers face is whether to assume either
ξ > 0 or ξ = 0, and apply the respective Fréchet or Gumbel distributions. One of the solutions
involves conducting a hypothesis test with the null hypothesis as:

H0 : ξ = 0

If there’s insufficient evidence to reject the null hypothesis (implying the tail index ξ is
insignificant), then then the Gumbel distribution should be used. If the tail index is significant (greater
than zero), the Fréchet distribution is most appropriate.

49
© 2014-2023 AnalystPrep.
Q.3996 Vanessa is trying to fit a distribution to the extreme loss tail of a historical financial return
dataset. After studying the data and consulting with end users, she has established the following:

I. T he loss data somewhat cluster; that is, losses are not strictly i.i.d.
II. T he end users do prefer that the extreme loss tail distribution itself exhibit right-skew; aka,
positive skew
III. T he distribution F (x) is actually unknown; i.e., it could be anything
Which of the established issues, in theory, effectively DISQUALIFIES the generalized
extreme-value (GEV) distribution as a candidate for application?

A. I only

B. II and III

C. III only

D. None of these issues disqualify the GEV distribution

T he correct answer is A.

For the generalized extreme value distribution to be used, the data set must be comprised of
independent, identically distributed random values. Preference for a positively skewed distribution is
in line with GEV’s two most common distributions - the Fréchet and Gumbel distributions.

50
© 2014-2023 AnalystPrep.
Q.3997 Which of the following statements about Extreme Value T heory (EVT ) and its application to
value at risk are true?

I. EVT extends the Central Limit T heorem to the distribution of the tails of independent,
identically distributed random variables drawn from an unknown distribution.
II. For empirical stock market data, the shape parameter in EVT is positive implying tails that
disappear more slowly than a normal distribution.
III. EVT can help avoid a shortcoming of the historical simulation method which may have
difficulty calculating VaR reliably due to a lack of data in the tails.

A. I only

B. II only

C. II and III only

D. All the above statements are correct.

T he correct answer is C.

I is incorrect. Extreme value theory is not governed by the central limit theorem because it deals
with the tail region of the relevant distribution.

II is correct. T he shape parameter in EVT for empirical stock market data is typically between 0.2

and 0.4, implying that the tails disappear more slowly than a normal distribution.

III is correct. Due to its reliance on historical data which may lack sufficient extreme values (i.e.,

extreme events), VaR calculation using the historical simulation method can be difficult; EVT can

help avoid this shortcoming.

51
© 2014-2023 AnalystPrep.
Q.3998 T he returns of a portfolio of stocks quoted in the S&P 500 share index over the last year
follow a distribution that is approximately normal. A trainee analyst removes some of the very best
performing stocks and produces reports based on the altered portfolio of returns. Which of the
following statements about the returns of the altered portfolio is/are correct?

I. T he distribution of returns of the altered portfolio is likely to be negatively skewed

II. T he distribution of returns of the altered portfolio is likely to be positively skewed
III. T he mean return is likely to be lower compared to the original portfolio.
IV. T he median return is likely to be higher compared to the original portfolio

A. II, III, and IV

B. I and II

C. II and III

D. I and III

T he correct answer is D.

T he distribution of returns is likely to be negatively skewed, since the extreme values on the
extreme right side of the distribution have been removed, therefore

I is correct

II is incorrect

Removing some of the highest values will decrease the mean and median, so III is correct and IV is

incorrect.

52
© 2014-2023 AnalystPrep.
Q.3999 Which of the following statements about Extreme Value T heory (EVT ) and its application to
value at risk is incorrect?

A. Unlike conventional approaches for estimating VAR, EVT considers only the tail behavior
of the distribution

B. Unlike conventional approaches for estimating VAR that assume that the distribution of
returns follows a unique distribution for the entire range of values, EVT allows the returns
to follow an unknown distribution.

C. EVT establishes the optimal point beyond which all values belong to the tail and then
separately models the distribution of tail events.

D. By segregating the tail of the distribution, EVT effectively ignores extreme events and
losses

T he correct answer is D.

A is correct. EVT uses only information in the tail.

B is correct. Conventional approaches such as delta-normal VAR and the historical method assume a

fixed probability density function (p.d.f.) for the entire distribution. T his could have the effect of

understating the extent of fat tails.

C is correct. EVT attempts to establish a cutoff point for the tail, and then estimates the parameters

of the tail distribution.

D is incorrect. EVT does not ignore extreme events (as long as they are in the sample). Rather,

emphasis is put on such events.

53
© 2014-2023 AnalystPrep.
Q.4000 For the standardized Gumbel, determine the 5% quantile, 95% quantile and the 99% quantile,
respectively.

A. -1.0972; 2.9702; 4.6001

B. -2.0845; 0.0052; 0.0041

C. 4.5614; 3.8542; 2.7823

D. -1.0965; 0.0052; -2.0485

T he correct answer is A.

First recall that for the standardized Gumbel, μ = 0, σ = 1

x = μ − σln[−ln (p)]

T he 5% quantile is : 0 − (1) ln [−ln(0.05)] = −1.0972

T he 95% quantile is : 0 − (1) ln [−ln(0.95)] = 2.9702

T he 99% quantile is : 0 − (1) ln [−ln(0.99)] = 4.6001

54
© 2014-2023 AnalystPrep.
Q.4001 In reality, natural and financial disasters are often related, and risk managers endeavor to have
some awareness of multivariate extreme risks. In this regard, why do risk managers limit themselves
to the study of a small number of financial variables at a time?

A. A lack of sufficient data

B. T he curse of dimensionality

C. Multivariate extremes are sufficiently rare that we need not worry about them

D. T he study of multivariate extremes is costly and time consuming.

T he correct answer is B.

Analysts study multivariate extreme events through EV copulas that have been developed over time.
In theory, a single copulas can have as many dimensions as the random variables of interest, but
there’s a curse of di mensi onal i ty in the sense that as the number of random variables considered
increases, the probability of a multivariate extreme event decreases.

For example, let’s start with the simple case where there are only two independent variables and

assume that univariate extreme events are those that occur only once for every 100 observations.

In these circumstances, we should expect to see one multivariate extreme event (both variables

taking extreme values) only one time in 1002; that’s one time in 10,000 observations. If we work

with three independent variables, we should expect to see a multivariate extreme event one time in

1003; that’s one time in 1,000,000 observations. T his clearly shows that as the dimensionality rises,

multivariate extreme events become rarer. T hat implies we have a smaller pool of multivariate

extreme events to work with, and more parameters to estimate. As such, analysts are forced to

work with a small number of variables at a time.

55
© 2014-2023 AnalystPrep.
Q.4002 In financial risk management, the clustering of high-severity risks is referred to as:

A. Multivariate risk analysis

B. Tail dependence

C. Univariate extreme value theory

D. Dimensional analysis

T he correct answer is B.

Tai l dependence refers to clustering of extreme events. Loosely speaking, tai l dependence
describes the limiting proportion that one variable exceeds a certain threshold given that the other
variable has already exceeded that threshold. In financial risk management, the clustering of high-
severity risks can have a devastating effect on the financial health of firms and this makes it an
important part of risk analysis.

Q.4003 In practice, risk analysts prefer the Peaks-over-threshold (POT ) approach over the
generalized extreme value approach because the POT approach:

A. Is less time consuming

B. Is more efficient in the use of data

C. Does not require the analyst to choose a threshold

D. All of the above

T he correct answer is B.

In practice, risk analysts prefer the POT approach over the GEV approach because the former is

more efficient in the use of data. In other words, the GEV approach involves some loss of useful data

relative to the POT approach, because some blocks might have more than one extreme in them. In

addition, the POT approach is better adapted to the risk measurement of tail losses because it

focuses on the distribution of exceedances over a threshold.

However, the POT approach comes with the problem of choosing a threshold, a situation that does

not arise with the GEV approach.

56
© 2014-2023 AnalystPrep.
Q.4004 To retrieve the value at risk (VaR) for the U.S stock market under the generalized extreme-
value (GEV) distribution, a risk analyst uses the following somewhat "realistic" parameters:

Location, μ = 4.0%

Scale, σ = 0.80

Tail index, ξ = 0.5

If the sample size, n = 100, then which is nearest to the implied 99.90% VaR?

A. 2.2500%

B. 7.4583%

C. 4.1002%

D. 2.5512%

T he correct answer is B.

Applying the Fréchet distribution (since ξ > 0), the VaR at α level of confidence is given by:

σn
VaR α = μn − [1 − (−nln (α))−ξn ]
ξn

At σ = 0.999,

0.8
VaR 0. 999 = 4 − [1 − (−100ln (0.999))−0. 5] = 7.4583%
0.5

57
© 2014-2023 AnalystPrep.
Q.4005 Assume the following observed parameter values:

β = 0.60

ξ = 0.30

μ = 1%

Nμ
n
= 5%

Compute the VaR at 99% confidence and the corresponding expected shortfall.

A. VaR = 2.25% ; ES = 3.5%

B. VaR = 2.2548% ; ES = 0.5252%

C. VaR = 2.2413% ; ES = 3.6304%

D. VaR = 1.5825% ; ES = 2.2385%

T he correct answer is C.

−ξ
β n
VaR = μ + [[ (1 − confidence level) ] − 1]
ξ Nμ
−0. 30
0.6 1
VaR 0. 99 = 1 + [[ (1 − 0.99)] − 1] = 2.2413%
0.3 0.05
VaR β − ξμ
ES = +
1− ξ 1 −ξ
2.2413 0.6 − 0.3 × 1
= + = 3.6304%
1 − 0.3 1 − 0.3

58
© 2014-2023 AnalystPrep.
Q.4006 Dyer and Blair Investment bank has an active position in commodity futures and is using the
peaks-over-threshold (POT ) approach (EVT ) to estimate value at risk (VaR) and expected shortfall
(ES) in accordance with extreme value theory. After careful consideration, the firm's risk managers
have settled on a threshold level of 5.00% to evaluate excess losses. T his choice of the threshold is
informed by the realization that 3.0% of the observations are in excess of this threshold value. As
displayed below, empirical analysis suggests the two other distributional parameters: scale, β = 0.70;
and shape (aka, tail index), ξ = 0.25.

Parameter Value
Loss threshold, u 5.00%
No. of observations, N 700
No. of observations that exceed threshold, N(u) 21
N(u)/N 3.00%
Scale,β 0.70
Shape, aka, tail,ξ 0.25

Determine the VaR at the 99% confidence level.

A. 0.02225

B. 0.04125

C. 0.05885

D. 0.05151

T he correct answer is C.

−ξ
β n
VaR = μ + [[ (1 − confidence level)] − 1]
ξ Nμ
−0. 25
0.7 1
VaR 0. 99 = 5 + [[ (1 − 0.99)] − 1] = 5.885%
0.25 0.03

59
© 2014-2023 AnalystPrep.
Q.4007 Kelvin Streetman is evaluating the extreme risks associated with active contracts on a
futures clearing house. He intends to use the peaks-over-threshold (POT ) approach (EVT ) to
estimate value at risk (VaR) and expected shortfall (ES) in accordance with extreme value theory.
Kelvin has set parameters at some empirically plausible values denominated in % terms as displayed
below:

Parameter Value
Loss threshold, u 5.00%
No. of observations, N 700
No. of observations that exceed threshold, N(u) 21
N(u)/N 3.00%
Scale,β 0.70
Shape, aka, tail,ξ 0.25

At the 99.0% confidence level, the position's VaR under the POT approach is 5.885%. Which is
nearest to the corresponding 99.0% expected shortfall (ES)?

A. 0.075426

B. 0.071133

C. 0.01885

D. 0.0225

T he correct answer is B.

VaR β − ξμ
ES = +
1−ξ 1−ξ
5.885 0.7 − 0.25 × 5
= + = 7.1133%
1 − 0.25 1 − 0.25

60
© 2014-2023 AnalystPrep.
Q.4008 Extreme value theory (EVT ) is a branch of applied statistics developed to address study and
predict the probabilities of extreme outcomes. Which of the following statements about EVT and its
applications is incorrect?

A. T he peaks-over-threshold approach provides the natural way to model exceedances over

a hi gh threshol d; which then determines the number of observed exceedances; the
threshold must be sufficiently high to apply the theory, but sufficiently low so that the
number of observed exceedances is a reliable estimate.

B. EVT estimates are subject to considerable model risk, and EVT results are often very
sensitive to the precise assumptions made.

C. Because observed data in the tails of distribution is limited, EV estimates can be very
sensitive to small sample effects and other biases

D. EVT asserts that distributions justified by the central limit theorem can be used for
extreme value estimation.

T he correct answer is D.

EVT differs from "central tendency" statistics where we seek to dissect probabilities of relatively
more common events, making use of the central limit theorem. Extreme value theory is not
governed by the central limit theorem because it deals with the tail region of the relevant
distribution.

EVT is subject to estimation risk, so B and C are correct.

Q.4009 According to the Fisher-T ippett theorem, as the sample size n gets large, the distribution of
extremes converges to:

A. a uniform distribution

B. a normal distribution

C. a generalized extreme value distribution

D. a generalized Pareto distribution

T he correct answer is C.

T he Fisher-T ippett theorem says that as the sample size n gets large, the distribution of extremes,
denoted M , converges to a generalized extreme value (GEV) distribution.

61
© 2014-2023 AnalystPrep.
Q.4010 As the threshold level, u, gets large, the Gnedenko–Pickands–Balkema–DeHaan (GPBdH)
theorem states that the distribution above-threshold observations converges to:

A. a normal distribution

B. a generalized extreme value distribution

C. a generalized Pareto distribution

D. a uniform distribution

T he correct answer is C.

T he second theorem of EVT is the Pickands-Balkema-de Haan theorem. It states that given a {X 1, ...,
X n} sequence of i.i.d. variables, the conditional distribution Fu(y) = P(X i - u < y|X i > u) of each
random variable X i of the sequence converges toward a generalized Pareto distribution for large u.

Q.4011 In setting the threshold in the POT approach, which of the following statements is most
accurate? Setting the threshold relatively low makes the model:

A. more applicable but decreases the number of observations in the modeling procedure.

B. less applicable and decreases the number of observations in the modeling procedure

C. more applicable but increases the number of observations in the modeling procedure.

D. less applicable but increases the number of observations in the modeling procedure.

T he correct answer is D.

T he GPD is considered the natural model for excess losses since all distributions of excess losses
converge to the GDP. To apply the GPD, however, we need to choose a reasonable threshold u,
which determines the number of observations, N μ , in excess of the threshold value. But the process
involves a trade-off. On the one hand, we have to come up with a threshold that is sufficiently high
for the GPBdH theorem to apply reasonably closely. On the other, we have to be careful not to
settle on a value that’s too high such that it leaves us with an insufficient number of excess-threshold
observations that lead to unreliable estimates.

62
© 2014-2023 AnalystPrep.
Q.4012 T he peaks-over-threshold approach generally requires:

A. fewer estimated parameters than the GEV approach and shares one parameter with the
GEV.

B. fewer estimated parameters than the GEV approach and does not share any parameters
with the GEV approach

C. more estimated parameters than the GEV approach and shares one parameter with the
GEV.

D. more estimated parameters than the GEV approach and does not share any parameters
with the GEV approach.

T he correct answer is A.

T he GEV has three parameters: location parameter μ, scale parameter σ, and shape (tail index)
parameter ξ

peaks-over-threshold approach, POT ( has only two parameters:

A positive scale parameter, β

A shape/tail index parameter, ξ that can be positive, zero, or negative. However, we limit

our interest to positive or zero values. (T his parameter is the same tail index encountered

under GEV theory.

63
© 2014-2023 AnalystPrep.
 Reading 66: Backtesting VaR

Q.1498 Value-at-risk (VaR) models are used for predicting risks; therefore, there must be a proper
process of validation which checks the adequacy of a model. T here are various tools and models to
check validation such as oversight, independent review, stress testing, and backtesting. Which of the
following statements are CORRECT about backtesting?
I. Backtesting is a statistical framework which provides verification that actual and projected losses
are in line
II. Backtesting compares the history of VaR forecasts to actual (realized) returns
III. Backtesting involves conducting reality checks which make up useful information for investment
decisions

A. Both I and II

B. Both II and III

C. Both I and III

D. All of the above

T he correct answer is A.

Statements I & II are correct. I. Backtesting is a statistical framework which provides verification

that actual and projected losses are in line. As such, backtesting compares the history of VaR

forecasts to actual (realized) returns.

Statement III is incorrect. T he reality checks provided by backtesting help risk managers in

checking their VaR forecasts but do not facilitate in investment decision-making.

64
© 2014-2023 AnalystPrep.
Q.1499 While conducting backtesting of VaR as an FRM manager, an accurate model is one where:

A. T he number of exceptions should be greater than the VaR significance level.

B. T he number of exceptions should be less than the VaR significance level.

C. T he number of exceptions should be equal to the VaR significance level.

D. None of the above.

T he correct answer is C.

For a model to be completely accurate, the number of exceptions would have to be the same as the
VaR significance level, where significance is 1 − Confidence level. Note that, backtesting period
constitutes a limited sample at a specific confidence level, which means it would be unrealistic to
expect to find the model-predicted number of exceptions in every sample. In other words, there are
instances where the observed number of exceptions will not be the same as that predicted by the
model, but that does not necessarily mean that the model is flawed. As such, we must establish the
level (point) at which we reject the model.

Q.1501 While carrying out backtesting of a leading bank’s VaR model, you have made the following
findings: the bank is currently calculating the 1-day VaR at a confidence level of 99%. However,
based on your findings you suggest changing the confidence level from 99% to 95%. Which of the
following statements would justify your stance?

A. While conducting backtesting with a 95% confidence interval, the probability of

committing a T ype 1 error is small as compared to the probability of T ype 1 error when
backtesting with 95% and 99% VaR models.

B. T he accuracy of the VaR model and the basis of accepting/rejecting the model have a
greater reliability at a 95% confidence level VaR as compared to at a 99% confidence level.

C. While conducting backtesting with a 95% confidence interval, the probability of rejecting
the VaR models at a 95% confidence level is equally to that at a 99% confidence level.

D. T here are fewer chances of a 95% VaR model being rejected based on backtesting as
compared to a 99% VaR model.

T he correct answer is B.

Using a lower VaR confidence level makes a bigger rejection region, giving room for more
exceptions. T his also makes the test more reliable.

65
© 2014-2023 AnalystPrep.
Q.1502 Matthew Hopkins is invited to interview for a position as a financial risk manager. After
completing an initial set of questions, the interviewer asks for the interpretation of the following
case: a $20 million 15-day VAR figure having a confidence level of 95%. Which of the following
represents the CORRECT interpretation?

A. T here is a 5 percent chance that there will be a gain of greater than $20 million in a time
period of 15 days.

B. T he corresponding VAR of the following day is $20 million, with a confidence interval of
95%.

C. T he amount of minimum loss spread over the next 15 days is at least $20 million with a
confidence of 95%.

D. T he amount of loss spread over the next 15 days is expected to be less than $20 million in
95 percent of case scenarios.

T he correct answer is D.

A 95% confidence level means that in 95% of the time, we expect losses not to exceed the stated
VaR amount. ($20 million in this case). Put differently, there's a 5% chance that over the next 15
days, there will be a loss exceeding $20 million.

66
© 2014-2023 AnalystPrep.
Q.1503 T he notion of backtesting generally includes the comparison of profits and losses on a daily
basis. T he 1996 Market Risk Amendment sheds light on the framework of backtesting. Which of the
following metrics are considered critical for such a process?
I. T he number of outliers
II. T he size of the outliers
III. T he risk measure to use

A. I and II

B. II and III

C. II only

D. None of the above

T he correct answer is A.

T he number and size of the outliers are significant measures. T he number of outliers is the number
of times when the realized loss exceeds the VaR value. Size refers to the extent to which outliers
have been grouped. T hat is, whether in time, they have become independent.
However, the model allows banks to use proprietary in-house models for measuring market risks.
Banks using proprietary models must compute VaR daily, using 99th percentile, one-tailed confidence
interval with a time horizon of ten trading days using a historical observation period of at least one
year.

67
© 2014-2023 AnalystPrep.
Q.1504 A detailed analysis of backtesting reveals that there are two major areas where backtesting
applies: in the calculation of the Value at Risk (VaR) and in the calculation of EPE1 profiles.
Considering this, the Basel Committee has stated very clear rules as to how to perform the VaR
backtest, along with factors distinguishing good and bad models. Which of the following statements
about the Basel Committee guidelines are CORRECT ?
I. T he Basel rules for backtesting are directly derived from the failure rate test.
II. As per Basel Committee, 5 exceptions are acceptable and fall under the green zone.
III. When exceptions fall in the red zone and become greater than or equal to 10, an automatic
penalty is generated.

A. I and II

B. II and III

C. I and III

D. All of them

T he correct answer is C.

Point number II mentions that 5 exceptions are acceptable, which is incorrect. Only 4 exceptions
are acceptable, and any number greater than this can result in a penalty.

Q.1505 It is a well-known fact that backtesting requires the application of quantitative, most often
statistical methods, for the purpose of determining whether a model for the assessment of risk is
adequate or not. Which of the following categories form part of the Basel Committee guidelines
regarding exceptions?
I. Model accuracy could be improved
II. Basic integrity of the model
III. Bad Luck and Intraday trading

A. I and II

B. II and III

C. I and III

D. All of them

T he correct answer is D.

All three options (A, B and C) show 4 categories used by the Basel Committee when assessing the
degree of penalty to be imposed on a party.

68
© 2014-2023 AnalystPrep.
Q.1508 T he underlying crux behind backtesting is that faulty models and bad luck must be separated,
or there must be a balance between T ype I and T ype II errors. However, backtesting has certain
limitations. Which of the following is NOT a limitation of backtesting?

A. Inability to model strategies affecting historic prices.

B. Potential curve fitting which means past successful strategies do not work in future.

C. Requirement of past conditions with adequate details.

D. Reducing volatility.

T he correct answer is D.

Reducing volatility is an advantage of backtesting and not a limitation.

Q.1509 Generally, the backtesting model focuses on unconditional coverage as it does not account
for time variations or conditioning in data. As a result, exceptions can bunch closely or cluster in
time. Which of the following backtesting outcomes does NOT raise a red flag?
I. At 95% confidence level, 12 exceptions occur on an annual basis and are spread evenly.
II. At 95% confidence level, 12 exceptions occur on an annual basis and 9 of these occurred over last
3 weeks.
III. At 95% confidence level, 12 exceptions occur on an annual basis - one exception per month.

A. I and II

B. II and III

C. I and III

D. All of them

T he correct answer is C.

If 9 out of 12 exceptions occurred over the last 3 weeks, this means there is increased volatility in
the market which is not captured by VaR. T his situation requires attention and raises a red flag.

69
© 2014-2023 AnalystPrep.
Q.2637 Jason Black, a risk analyst at a large multinational bank, is backtesting the VaR model of the
bank. T he model being tested is a daily, 98% VaR model. If the backtest is conducted for one year at a
95% confidence level, what is the acceptable number of daily losses that will lead Black to conclude
that the model is calibrated correctly?

A. 6

B. 9

C. 12

D. 5

T he correct answer is B.

Since the test is being conducted at a 95% level, the cutoff value for the test will be 1.645. To test
whether the model is accurate, the following hypothesis is tested:

(x − pT )
> z = 1.645
√[p(1 − p)T ]
x − 0.02 × 252
= 1.645
√0.02 × 0.98 × 252
⇒ x = 8.895

T he acceptable number of exceptions allowed for Black to conclude that the model is correctly
calibrated is 9.

70
© 2014-2023 AnalystPrep.
Q.2638 A model gives a VaR value of $9.5 million for a portfolio at a 99% confidence interval. A one-
year backtest conducted at the 95% confidence level reveals that losses exceeded $9.5 million on 12
occasions. T he model is accepted as accurate. Assuming 224 days in a year, which of these
statements is most likely true?

A. A T ype I error has occurred.

B. A T ype II error has occurred.

C. Both T ype I and T ype II errors have occurred.

D. T he model has been accepted correctly.

T he correct answer is B.

To test whether the model is accurate, the following hypothesis is tested:

(x − pT )
> z = 1.65
√[p(1 − p)T ]
12 − 0.01 × 224
= 1.65
√0.01 × 0.99 × 224
⇒ x = 6.55

T he z-value calculated is 6.55, which is greater than 1.65. T hus, the hypothesis that the model is
correctly calibrated can be rejected. Accepting an inaccurate model is a type II error. T herefore, it
can be said that the model is not accurate. Accepting an inaccurate model is a T ype II error.

71
© 2014-2023 AnalystPrep.
Q.2640 Willy Jones and Craig Atherton are two junior risk analysts. T hey have recently been
assigned to perform a 1-year backtest of a 1 day 98% VaR model, assuming 225 days in the year.
During the next few days, they exchange a number of emails regarding the assignment:

Email 1 - Jones forecasts the number of expected exceptions for the model to be 4.5.

Email 2 - Atherton replies that according to the Basel Committee’s prescribed penalty zones, the
yellow zone starts at six exceptions and attracts a multiplier of four.

Email 3 - Jones states a type II error occurs when an accurate model is rejected.

T he contents of which of the emails is/are not true?

A. Emails 1 and 2

B. Emails 2 and 3

C. Emails 1 and 3

D. All three emails

T he correct answer is B.

Both statements 2 and 3 are incorrect.

T he number of forecasted exceptions are (1 - 0.98) × 225 = 4.5

According to the Basel Committee, the yellow zone is between 5 and 9 exceptions and attracts a
multiplier between 3.4 and 3.85.

Rejecting an accurate model is classified as a type I error. A type II error occurs when an inaccurate
model is accepted.

72
© 2014-2023 AnalystPrep.
Q.2641 What is the number of exceptions that are forecasted during the backtesting of a VaR model
that is constructed using a 95% confidence interval over a 1000-day period?

A. 100

B. 10

C. 25

D. 50

T he correct answer is D.

T he number of exceptions is equal to 1 - confidence level or 1 - 0.95 = 5%. Multiplying this by the

number of days used will give us the exact number of exceptions:

5% × 1000 = 50

Q.2642 According to the Basel Committee rules for the backtesting of VaR, which of the following
statements in relation to the number of exceptions and the corresponding capital multiplier is NOT
accurate?

A. If the number of exceptions is between 5 and 9, the bank will fall in the yellow zone.

B. If the number of exceptions is between 5 and 9, a capital multiplier of 3 will be applied.

C. If the number of exceptions exceeds 10, a capital multiplier of 4 will be applied.

D. Banks in the red zone will be charged the highest penalty.

T he correct answer is B.

T he Basel Committee has defined penalty zones based on the number of exceptions, which are as
follows:

Zone Number of Exceptions Muliplicative Factor

Green 0 to 4 3.0
Yellow 5 3.4
6 3.5
7 3.65
8 3.75
9 3.85
Red 10+ 4.0

73
© 2014-2023 AnalystPrep.
Q.2643 T he Basel Committee has defined four maj or reasons for exceptions found during
backtesting. T hese include all of the following, except:

A. Model not calibrated to market conditions

B. Model lacks basic integrity

C. Intraday trading

D. Bad luck

T he correct answer is A.

According to the Basel Committee rules for backtesting of VaR, four different categories have been
defined for reasons of exceptions. T hese include the need for improvement in model accuracy, lack
of integrity of the model, exceptions caused by intraday trading activity and lastly bad luck or cases
where market conditions varied significantly from the norm.

74
© 2014-2023 AnalystPrep.
Q.2824 T he management of a financial institution reports that on a particular year, the daily revenue
fell short of the downside 95% VaR band on 24 occasions (days), or more than 5% of the time. Ten of
these 24 occurrences fell within the May to July period. Assuming 252 days in the year, find the test-
statistic and test if this was a faulty model or bad luck using the binomial distribution.

A. 8.32; it is a faulty model

B. 2.16; it is bad luck

C. 1.01; it is bad luck

D. 3.3; it is a faulty model

T he correct answer is D.

Considering the equation:

(x − pT )
Z= ≈ N (0, 1)
√p(1 − p)T

we have x = 24, p = 5% , T = 252.

Applying these values in the equation, we get:

(24 − 0.05 × 252)

Z=
√0.05(1 − 0.05) × 252
= 3.3

Which is larger than the cut off value of 1.96.

T herefore, we reject the hypothesis that the VaR model is unbiased. It is not likely that this was bad

luck at the 95% confidence level.

75
© 2014-2023 AnalystPrep.
 Reading 67: VaR Mapping

Q.1511 Risk measurement is widely practiced in the financial sector to establish the risk
characteristics of trading instruments and portfolios. T his is done via several methods some of which
can be time-consuming and complex because it is literally impractical to measure all risk factors
individually. Hence, the VaR method is used through the process of mapping to:

A. Simplify a portfolio by mapping positions on selected risk factors

B. Simplify a portfolio by mapping positions on all risk factors which can have a minor or
major impact on the performance of an instrument

C. Simplify portfolio by mapping positions on only five risk factors

D. Simplify portfolio by mapping positions on all abnormal risk factors which can impact the
performance of an instrument

T he correct answer is A.

Whichever value-at-risk (VaR) method is used, the risk measurement process needs to simplify the
portfolio by mapping the positions on the selected risk factors. It would be too complex and time-
consuming to model all positions individually as risk factors.

Q.1512 Mapping is a useful process and also an instructive one because it provides useful judgments
about risk drivers of derivatives. Financial institutions cannot always use historical prices to develop
a risk profile for the instrument. In addition, they cannot develop risk profiles of options on the basis
of historic values. T herefore, mapping gives us a way to handle these practical problems when:

A. characteristics of instrument do not change over time.

B. characteristics of instrument change over time.

C. a large number of factors needs to be measured separately for each position.

D. the characteristics of the instrument are only exposed to a single major risk factor.

T he correct answer is B.

Mapping is the only solution when the characteristics of the instrument change over time. T he risk
profile of bonds, for instance, changes as they age. T he bonds must be mapped on yields that best
represent their current profiles. Similarly, the risk profile of options changes very quickly. Options
must be mapped on their primary risk factors. Mapping provides a way to tackle these practical
problems.

76
© 2014-2023 AnalystPrep.
Q.1513 As we know, mapping is the simple process of analyzing the market instruments on primitive
risk factors. Considering this concept, let’s take a single instrument that has a market value of Vm . It
is allocated to specific risk exposures namely X 1, X 2, and X 3. Suppose not all of the current market
value Vm is allocated to these risk factors. What does that imply with regard to the remaining value?

A. T he remainder value is not exposed to any risk.

B. T he remainder value is allocated to cash which is not a risk factor.

C. T he remainder value is allocated to a separate set of risk factors.

D. T he remainder value’s risk exposure is very difficult to measure.

T he correct answer is B.

If not fully allocated to the risk factors, it must mean that the remainder value is allocated to cash
which is not a risk factor due to the fact that it has no risk.

Q.1514 In the process of mapping, finance experts first need to choose the most effective set of
primitive risk factors against which the market instrument will be positioned to measure risk. T his
choice of factors must be balanced between time devotion and accurate risk measurement. In short,
the choice of primitive risk factors should reflect:

A. T he easiest way to get better results in the least amount of time.

B. T he trade-off between models with a large number of factors and less complex models.

C. T he trade-off between better quality of approximation and faster processing.

D. T he trade-off between specific risks with significant effects and those with insignificant
effects.

T he correct answer is C.

T he choice should reflect the trade-off between better quality approximation and faster processing.
More factors lead to tighter risk measurement but also require more time devoted to the modeling
process and risk computation.

77
© 2014-2023 AnalystPrep.
Q.1515 Once we have selected the risk factors, then, the next step is to map the portfolio or
instrument positions against these risk factors which can be achieved through any of the three
approaches of mapping, depending on the best suitable approach. In choosing the mapping approach,
which important factor should be kept in mind?

A. Mapping should only preserve the market value of the instrument.

B. Mapping should preserve the par value as well as the market risk of the position.

C. Mapping should preserve the market value as well as the interest rate risk of the position.

D. Mapping should preserve the market value as well as the market risk of the position.

T he correct answer is D.

T he essence of mapping is to identify risk factors that “explain” the current values of the portfolio
positions. By preserving the market value, you enforce the assumption that the total value of the
position is attri butabl e to the risk factors identified. By preserving the market risk, you enforce
the assumption that the risk factors identified are a true representati ve of total market risk.

Q.1516 Considering an example of a two-bond portfolio, we calculated the portfolio returns and the
risks associated with those portfolios using the mapping technique. T hen, we found some specific
values, say, 2.80 VaR for duration mapping and 2.67 VaR for cash flow mapping. T his notable
difference in these values is due to the fact that:

A. risk measures are not perfectly linear with maturity and correlations are below unity.

B. risk measures are perfectly linear with maturity and correlations are below unity.

C. risk measures are perfectly linear with maturity and correlations are above unity.

D. risk measures are not perfectly linear with maturity and correlations are equal to unity.

T he correct answer is A.

T he difference is due to two factors. First, risk measures are not perfectly linear with maturity.
Secondly, correlations are below unity, which reduces risk even further.

78
© 2014-2023 AnalystPrep.
Q.1517 Benchmarking is the process of evaluating a portfolio’s risk against some standard or ideal
portfolio risk that is considered as the benchmark. T herefore, the VaR of the deviation of portfolio A
relative to the benchmark is

T racking Error VaR = α √(x − x 0 )′ Σ(x– x 0 )

After we performed the necessary calculations for portfolio A, we found the tracking error VaR of
portfolio A which is 0.63 [Link] does this tracking error VaR value imply?

A. T he maximum deviation between the index and portfolio A is at most 0.63 million under
normal market conditions.

B. T he minimum deviation between the index and portfolio A is at most 0.63 million under
normal market conditions.

C. T he maximum deviation between the index and portfolio A is at most 0.63 million under
abnormal market conditions.

D. T he minimum deviation between the index and portfolio A is at most 0.63 million under
abnormal market conditions.

T he correct answer is A.

T he tracking error VaR (T E-VaR) is the maximum deviation between the index and the portfolio
under normal market conditions.

79
© 2014-2023 AnalystPrep.
Q.1518 T he stress testing approach of assessing risk exposure represents the link between
calculating VaR through matrix multiplication and movement in underlying prices. T his test gives
more direct and appropriate results. Keeping this in mind, which of the following is true about stress
testing as described above?

A. Stress testing is used to evaluate the potential impact on portfolio values of unlikely
events or movements in a set of financial variables.

B. Stress testing is a risk management tool that compares predicted results to observed
actual results (historical data).

C. Both A and B options are true.

D. None of the above are true.

T he correct answer is A.

Stress testing is an approach which computes VaR through matrix multiplication and through
movements in underlying prices. Computing VaR through matrix multiplication is much more direct
and more appropriate because it cuts across different sectors of the yield curve.

Q.1519 Forward contracts are the simplest types of derivatives and their risk can easily be
calculated through basic building blocks forming those contracts. But before buying forward
contracts, an investor needs to make a decision between two alternatives which are economically
equivalent. T he usual options available to the investor are to:

A. buy X units of any asset at price P and sell them at a higher price to potentially earn
profits or enter into a forward contract to buy one unit of the asset in one period.

B. buy X units of any asset at price P and hold them for one period or enter into a forward
contract to buy one unit of the asset in two periods.

C. buy X units of any asset at price P and hold them for one period or enter into a forward
contract to buy one unit of the asset in one period.

D. buy X units of any asset at price P and sell them at a higher price to potentially earn
profits or enter into a forward contract to buy one unit of the asset in one period at the
lowest price possible.

T he correct answer is C.

Investors have two alternatives that are economically equivalent: (1) Buy X units of the asset at the
prevailing market prices and hold them for one period, or (2) enter into a forward contract to buy
one unit of the asset in one period.

80
© 2014-2023 AnalystPrep.
Q.1520 One of the methods of cash flow mapping involves decomposing bond risk into the risk of
each of the bond's cash flows. T his describes:

A. Principal mapping

B. Duration mapping

C. Cash flow mapping

D. Present value mapping

T he correct answer is C.

Under cash flow mapping, the risk of the bond is decomposed into the risk of each of the bonds' cash

flows. T he present value of cash flows is mapped onto the risk factors for zeros of the same

maturity.

Under principal mapping, we only consider the risk of the repayment of the principal amount.

Under duration mapping, the risk of the bond is mapped to a zero-coupon bond of the same duration.

“Present value mapping” is nonexistent.

81
© 2014-2023 AnalystPrep.
Q.1522 To determine the value of forward contracts, we decompose the forward contract into its
main building blocks which will impose the net effect on the risk and price of the forward contract.
T his methodology can also be used for long-term currency swaps which are typically identical to
portfolios of forward contracts. Keeping this scenario in mind, which of the following statements is
true?

A. A 5-year contract to pay dollars and receive Euros is equivalent to a series of 5 forward
contracts to exchange a set amount of dollars.

B. A 5-year contract to pay dollars and receive Euros is equivalent to a series of any number
of forward contracts to exchange a set amount of dollars.

C. A 5-year contract to pay dollars and receive Euros is equivalent to a forward contract to
exchange a set amount of dollars.

D. A 5-year contract to pay dollars and receive Euros is not equivalent to a series of 5
forward contracts to exchange a set amount of dollars.

T he correct answer is A.

A swap contract is equivalent to a portfolio of forward contracts with identical delivery prices and
different maturities. Consequently, swap contracts are similar to forwards in that (1) at any date,
swap contracts can have positive, negative, or no value, and (2) at initiation, the fixed amount paid is
chosen so that the swap contract is costless. T he unique fixed amount which zeros out the value of a
swap contract is called the swap price.

82
© 2014-2023 AnalystPrep.
Q.1523 T he valuation of commodity forward contracts is much more complex compared to that of
financial assets such as currencies or stock indices because these commodity-based contracts do not
have well-defined income flows and most of the time do not make monetary payments. Rather, items
are consumed giving an implied benefit called convenience yield, which represents:

A. the quantifiable disadvantage to owning the commodity rather than buying the futures
contract.

B. the quantifiable advantage to owning the commodity rather than buying the futures
contract.

C. the addition of the risk-free rate and the storage cost.

D. the cost of storage cost from holding the commodity.

T he correct answer is B.

T he flow of benefits net of storage cost is loosely called convenience yield to represent the benefits

from holding the cash product. In other words, the convenience yield is an implied return on holding

inventories. It is an adjustment to the cost of carry in the non-arbitrage pricing formula for forward

prices in markets with trading constraints.

Mathematically, it can be written as:

S0 e(r+λ−c)

Where

S0 is the spot price;

r is the risk-free rate of return;

λ is the storage cost; and

c is the convenience yield.

We can see from this that the convenience yield lowers the price of the contract because it is the

quantifiable advantage to owning the commodity rather than buying the futures contract. For

example, I have wheat in case I need it to eat, rather than simply having a wheat futures contract.

83
© 2014-2023 AnalystPrep.
Q.1524 A forward rate agreement is a type of forward contracts that allows the contracting parties
to make transactions in a locked interest rate at some future date. T he buyer in this contract locks
in the borrowing rate and the seller locks in the lending rate at some future date.
Which statement is true about such a contract if the spot rate is higher than the forward rate at the
time of the transaction?

A. T he buyer will be worse off and will receive payments at a lower rate and the seller will
also be worse off by lending at a lower rate.

B. T he buyer will be worse off and will receive payments at a lower rate while the seller
will benefit by lending at a lower rate.

C. T he buyer will benefit and will receive payments at a lower rate while the seller will be
worse off by lending at a lower rate.

D. Both the buyer and the seller will be in the same position with no effect in benefits and
losses.

T he correct answer is C.

Forward rate agreements (FRAs) are forward contracts that allow users to lock in an interest rate at

some future date. T he buyer of an FRA locks in a borrowing rate and the seller locks in a lending

rate. In other words, the “long” receives payments if the spot rate is above the forward rate.

It is worth noting that, FRA usually involves two parties exchanging a fixed interest rate for a

variable interest rate. T he party that pays the fixed rate is referred to as the borrower, on the other

hand, the party that is pays the variable rate is referred to as the lender.

A borrower enters into an FRA believing that the interest rates will rise and that he will lock in an

interest rate if in the future.

84
© 2014-2023 AnalystPrep.
Q.1525 T he forward rate can be defined as the implied rate that makes the return on a T 2 period
investment and a T 1 period investment equal. T hat is:

(1 + R 2 T 2) = (1 + R 1 T 1)[1 + F1,2(T 2 − T 1 )]

It means that if you sold a 5*10 FRA on $50 million, this transaction is equal to borrowing $50 million
into 5-month Bills and investing the proceeds into 10-month [Link] of the following formulas is
supporting the above statement?

A. Long 6*12FRA = Short 6-month Bills + Long 12-month Bills

B. Long 6*12FRA = Short 12-month Bills - Long 6-month Bills

C. Long 6*12FRA = Long 12-month Bills + Short 6-month Bills

D. Long 6*12FRA = Long 6-month Bills + Short 12-month Bills

T he correct answer is D.

Long 6 ∗ 12F RA = Long 6 − month Bill + short 12 − month bill

Q.1526 Interest rate swaps are the most commonly used derivatives because of their less volatile
risk positions. An interest rate swap is an agreement between two parties to exchange interest rate
flows on the basis of fixed to floating rates and vice versa. T hey can be broken down into two legs: a
fixed leg and floating leg. T he fixed leg can be the price on a:

A. floating-rate note and the floating leg can be equivalent to a coupon-paying bond.

B. coupon-paying bond and the floating leg can be equivalent to a floating-rate note.

C. zero-coupon bond and the floating leg can be equivalent to a floating-paying bond.

D. floating-paying bond and the floating leg can be equivalent to a zero-coupon bond.

T he correct answer is B.

Swaps can be decomposed into two legs, a fixed leg and a floating leg. T he fixed leg can be priced as a
coupon-paying bond and the floating leg is equivalent to a floating-rate note (FRN).

85
© 2014-2023 AnalystPrep.
Q.1527 Risk measurement is difficult for non-linear derivatives or options because of non-linearity.
To simplify the process, the Black-Scholes model is used.
What is the assumption of this model other than perfect capital markets?

A. Underlying spot prices follow a continuous geometric Brownian motion with constant
volatility.

B. Underlying spot rates follow a continuous algebraic Brownian motion with constant
volatility.

C. Underlying spot prices follow a stationary geometric Brownian motion with constant
volatility.

D. T here is no other assumption of this model except for perfect capital markets.

T he correct answer is A.

T he BS model assumes, in addition to perfect capital markets, that the underlying spot price follows a
continuous geometric Brownian motion with constant volatility.

Q.1528 Consider the Black-Scholes (BS) model for European options. Suppose we drew a graph
showing the relationship between delta (the first partial derivative of a nonlinear option) and spot
prices of options with differing maturities. What would be the relationship observed for long-term
and short-term options?
I. T he relationship becomes more nonlinear for short-term options than long-term options
II. T he relationship becomes more linear for short-term options than long-term options
III. Linear approximations may be acceptable for options with long maturities when the risk horizon
is short

A. I and III only

B. II and IIII only

C. II only

D. III only

T he correct answer is A.

Delta increases with the underlying spot price and the relationship becomes more nonlinear for
short-term options. Secondly, linear approximations may be acceptable for options with long
maturities when the risk horizon is short.
Further explanation:
As stated in the chapter:
"T he figure shows that delta is not a constant, which may make linear methods inappropriate for
measuring the risk of options. Delta increases with the underlying spot price. T he relationship

86
© 2014-2023 AnalystPrep.
becomes more nonlinear for short-term options, for example, with an option maturity of 10 days.
Linear methods approximate delta by a constant value over the risk horizon. T he quality of this
approximation depends on parameter values."
...
"T hus linear approximations may be acceptable for options with long maturities when the risk
horizon is short."

Further explanation:
T he delta for short-term options changes faster than the delta of long-term options because as the
time to expiration decreases, it means that any change in the underlying's price is unlikely to
evaporate fast enough, and will therefore most likely persist up until the maturity date and impact
the decision whether or not to exercise the option. If the change is likely to persist, then the option
price (and hence delta) will also change faster and actually exhibit more of a "jump", a non-linear
change.
In fact, as the time remaining to expiration grows shorter, the time value of the option evaporates
and correspondingly, the delta of in-the-money (IT M) options increases faster relative to longer-term
IT M options. Similarly, the delta of out-of-the-money (OT M) options decreases faster relative to that
of longer-dated OT M options.

87
© 2014-2023 AnalystPrep.
Q.1529 Risk measurement should always be a prioritized endeavor for financial institutions. In this
regard, financial instruments need to be mapped on a set of primitive risk factors. T he art of risk
management lies in the ability to choose an appropriate set of risk factors. Keeping this in mind,
which of the following statements is/are true?
I. A large number of risk factors should be incorporated to avoid any future loses
II. Only a few risk factors should be selected to save time and make decisions in a timely fashion
III. T here should be proper allocation of primitive risk factors to avoid slow and wasteful
measurements
IV. T here should be a proper set of general market and specific risk factors depending on the position
of the instrument

A. I and IV

B. III and IV

C. IV only

D. I and III

T he correct answer is B.

Too many risk factors would be unnecessary, slow, and wasteful. Too few risk factors, in contrast,
could create blind spots in the risk measurement system.

88
© 2014-2023 AnalystPrep.
Q.1530 In practice, we have to keep the number of risk factors small during mapping. T hese risk
factors include both general market risks and specific market risks for the entire portfolio. T hus the
portfolio return is calculated including variance through the following equation on:

n
V (R P ) = (B2p )V (R m ) + ∑ (W i2 )(σgi)
i=1

T his decomposition shows that:

A. with less general market risk factors, there will be less specific risk factors for fixed
amount of total risk, (V )(R p) .

B. with more general market risk factors, there will be more specific risk factors for fixed
amount of total risk, (V )(R p) .

C. there will be equal general market and specific risk factors for a fixed amount of total
risk, (V )(R p) .

D. with more general market risk factors there will be less specific risk factors for fixed
amount of total risk, (V )(R p) .

T he correct answer is D.

T his decomposition shows that with more detail on the primitive or general-market risk factors,
there will be less specific risk for a fixed amount of total risk V (R p) .

Q.2644 Adding more general risk factors to a VaR model will most likely:

A. Increase the size of specific risks.

B. Decrease the size of specific risks.

C. Have no effect on the size of specific risks.

D. None of the above. While adding more factors to a model will affect the size of specific
risk it is not possible to determine the exact nature of the change without considering what
factors are added.

T he correct answer is B.

T he number of general risk factors used in a VaR model will affect the size of the specific or residual
risk. For instance, a model that only considers the duration of the portfolio will have a larger
residual/specific risk as compared to a model that considers both duration and credit risk. Risk can be
better defined by adding more risk factors. Adding general risk factors to the model will help define
the risk more accurately and will reduce the specific or residual risk.

89
© 2014-2023 AnalystPrep.
Q.2645 All of the following are VAR mapping systems for fixed-income securities, except:

A. Principal mapping

B. Duration mapping

C. Cash flow mapping

D. Interest mapping

T he correct answer is D.

T here are three different VaR mapping approaches. T hese include principal mapping, duration
mapping, and cash-flow mapping.

90
© 2014-2023 AnalystPrep.
Q.2825 Determine the forward rate for a 1-year forward contract to exchange US dollars for Euros.
It is estimated that the Euro spot is $1.3988. T he 1-year EURO T-bill is quoted at 2.28% while the 1-
year USD T-bill is quoted at 3.33%.

A. 1.41 USD/EURO

B. 1.31 USD/EURO

C. 1.22 USD/EURO

D. 1.5 USD/EURO

T he correct answer is A.

(1 + AFd × rd )
F X f wd = F X s pot × [ ]
(1 + AFf × rf )

Where:

F X f wd = fair forward rate (units of domestic currency / unit of foreign currency)

F X s pot = spot FX rate (units of domestic currency / unit of foreign currency)

rd = domestic annual interest rate

rf = foreign annual interest rate

AFd = domestic accrual factor (days of accrued interest / days in a year)

AFf = foreign accrual factor

T hus,

360
⎡ (1 + 360×0. 033) ⎤
F X f wd = 1.3988 × = 1.41 USD/EURO
⎣ (1 + 360 ⎦
)
360×0. 0228

91
© 2014-2023 AnalystPrep.
Q.2826 A bank has a cash flow decomposition with a duration of 5 years. Given that the VaR of the
index at the 95% confidence level is $2.080 million, with a tracking error of $1.09 million, calculate
the variance improvement relative to the original index.

A. 23.5%

B. 33.6%

C. 72.5%

D. 95.1%

T he correct answer is C.

Recall that the variance improvement is given by:

T racking Error VaR 2 1.09 2

1 −( ) =1 −( ) = 0.725 = 72.5%
Absolute risk of the index 2.08

Note:

To benchmark a portfolio, we measure the VaR of the portfolio relative to the VaR of a benchmark.

T he VaR of the deviation between the two portfolios is referred to as a tracking error VaR. T he

difference comes in because it is possible to construct portfolios that match the risk factors of a

benchmark portfolio but have either a higher or a lower VaR.

If x is the vector position of the portfolio and x 0 the vector position of the index, then the tracking
error VaR is given by:

T racking Error VaR = α √(x— x 0)′ ∑(x— x 0)

Now, once we have the tracking error VaR, and if the absolute risk of the index is known, we can
now go ahead to calculate the variance improvement as we have done above.

92
© 2014-2023 AnalystPrep.
Q.2827 T he following table gives VaR percentages at the 95% confidence level for a bond with
matuities ranging from one year to 5 years:

Maturity VaR
1 0.4777
2 0.9961
3 1.4264
4 1.9618
5 2.4120

A bond portfolio consists of a $100 million bond maturing in one year and a $100 million bond
maturing in three years. Determine the VaR of this bond portfolio using the principal VaR mapping
method.

A. $1.2235m

B. $1.7765m

C. $1.9922m

D. $1.5m

T he correct answer is C.

(1+3)
T he VaR percentage is 0.9961 for a two-year zero-coupon bond [ = 2] .
2

Principal mapping

VaR = VaR percentage × market value of the average life of the bond

Principal mapping

V aR = 0.9961% × $200m = $1.9922m

93
© 2014-2023 AnalystPrep.
Q.2828 Calculate the current forward rate if you are given that the spot price of 1 unit of the
underlying cash asset is 1.22, with a domestic free rate of 0.037 and τ = 1. T he income flow rate y is
1.92%. (Assume continuous compounding.)

A. 1.24

B. 0.78

C. 1.32

D. 1.50

T he correct answer is A.

Remember that the current forward rate is given by the equation:

Ft = St e−ytert

We know that, St = 1.22, r = 0.037, y = 0.0192 and τ = 1.

Applying the formula:

Ft = 1.22e−0. 0192×1 × e0. 037×1 = 1.24

T he forward rate to but one unit of the underlying cash asset is 1.24

Q.3015 Which of the following components is NOT a relevant factor while calculating the total VaR
of a USD corporate bond for a US investor?

A. T he USD swap interest rate

B. T he bond sector's credit spreads

C. T he maturity of the bond

D. FX rates

T he correct answer is D.

For a US investor, FOREX rates are not relevant for VaR calculation if the bond in question is
denominated in USD. All the other 3 factors are necessary inputs.

94
© 2014-2023 AnalystPrep.
Q.3016 Which of the following inputs is NOT required in order to calculate the 99% Monte Carlo 1-
day VaR of a portfolio made of two stocks A and B, assuming both stocks have normally distributed
returns?

A. T he correlation of the returns between A and B

B. T he credit rating of entities A and B

C. T he spot values of stocks A and B

D. Normally distributed random numbers

T he correct answer is B.

Credit ratings are irrelevant for a market risk VaR calculation of both stocks.

95
© 2014-2023 AnalystPrep.
Q.3040 Mapping refers to the process of replacing the current values of a portfolio with risk factor
exposures. More generally, it is the process of replacing each instrument by its exposures on
selected risk factors. Mapping is important because:

A. it helps us to cut down on the dimensionality of covariance matrices and correlations.

B. it helps avoid rank correlation problems.

C. it greatly reduces the time needed to carry out risk assessment and related calculations.

D. All of the above.

T he correct answer is D.

Mapping helps us to cut down on the dimensionality of covariance matrices and correlations.

Given a portfolio comprising of n instruments, we would need to gather data on n volatilities and n(n-

1)/2 correlations, resulting in a labyrinth of pieces of information. As n increases, so does the amount

of information we have to collect and process. It is important to keep the dimensionality of our

covariance matrix at a manageable level to avoid computational problems.

Mapping helps avoid rank correlation problems

By handling a large number of risk factors that are closely correlated (or even perfectly correlated

in extreme cases), we might run into rank problems with the covariance matrix and end up producing

pathological estimates that might lead to erroneous conclusions. To avoid such problems, it is

important that we select an appropriate set of risk factors that are not closely related.

Mapping greatly reduces the time needed to carry out risk assessment and related calculations. By

reducing a portfolio comprised of a large number of different positions to a consolidated set of risk-

equivalent positions in basic risk factors, it is possible to conduct calculations at a faster speed. T he

only downside to such a move is that precision is lost.

96
© 2014-2023 AnalystPrep.
Q.3041 If portfolio assets are perfectly correlated, portfolio VaR will equal:

A. Component VaR

B. Marginal VaR

C. Diversified VaR

D. Undiversified VaR

T he correct answer is D.

If we assume perfect correlation among assets, VaR would be equal to undiversified VaR.

97
© 2014-2023 AnalystPrep.
 Reading 68: Messages from the Academic Literature on Risk
Measurement for the Trading Book

Q.1531 One of the fundamental issues when using VaR for regulatory capital is the horizon over
which VaR is calculated. Scaling up the short-horizon VaR to the desired time period using the square-
root-of-time may compromise the accuracy of VaR because:

A. the tail risk is likely to be overestimated.

B. the downward bias tends to decrease as we increase the time horizon.

C. the tail risk is likely to be underestimated.

D. the upward bias tends to decrease as we increase the time horizon.

T he correct answer is C.

T he square root of time has been discredited for the following reasons:

For starters, the rule assumes that risk factors are i.i.d. i.e., independent and normally

distributed. In other words, the square root of time rule assumes that price moves are

independent. In practice, the distribution of losses changes with time. T here are time

periods with more losses than others.

When applied to quantiles, the square root of time rule assumes that returns follow a

normal distribution. In reality, returns exhibit excess kurtosis, e.g. they are “fat-tailed.”

It has also been established that when risk factors exhibit jumps (sudden and robust loss-

generating moves), scaling by the square root of time systematically under-estimates risk.

In fact, the downward bias tends to be amplified by longer horizons. T he argument is that

the rule does not sufficiently scale jump risk.

As a result, even if the square-root-of-time rule has widespread applications in the Basel Accords, it

fails to address the objective of the Accords.

98
© 2014-2023 AnalystPrep.
Q.1532 T he intra-horizon VaR is a risk measure that combines VaR over the regulatory horizon with
P&L fluctuations over the short term. Taking intra-horizon risk into account generates risk
measures consistently higher than standard VaR, up to multiples of VaR, and:

A. the divergence is larger for derivative exposures.

B. the minimum cumulative loss exerts a distinct effect on the capital of a financial
institution.

C. the divergence is smaller for derivative exposures.

D. the information is carried on low-frequency P&L.

T he correct answer is A.

Bakshi and Panayotov (2010) found that taking intra-horizon risk into account generates risk
measures consistently higher than standard VaR, and the divergence is larger for derivative
exposures.

Q.1533 According to academic literature, “time-varying volatility in financial risk factors is important
to the VaR.” When the true underlying risk factors exhibit time-varying volatility, the use of
historically simulated VaR without incorporating time-varying volatility can:

A. Reduce pro-cyclicality

B. Under-estimate risk

C. Increase instability

D. Over-estimate risk

T he correct answer is B.

Option B is the correct answer because Risk Metrics Technical document showed theoretically that
using historical simulation VaR without incorporating time-varying volatility can dangerously under-
estimate risk.

99
© 2014-2023 AnalystPrep.
Q.1534 Multivariate GARCH models such as the BEKK model of Engle and Kroner (1995) or the DCC
model of Engle (2002) can be used to estimate:
I. T ime-varying volatilities
II. Correlations
III. Large numbers of positions

A. I and II

B. II and III

C. I, II and III

D. None of the above

T he correct answer is C.

Multivariate GARCH models can be used to estimate time-varying volatilities as well as correlations.
In addition, some recent advances in literature allow us to estimate a multivariate GARCH-type
model when there is a large number of risk factors.

Q.1535 T he amalgamation of VaR models and market liquidity requires a distinction between
exogenous and endogenous liquidity. Which of the following descriptions is correct?

A. T he endogenous component of liquidity risk corresponds to the average transaction costs

set by the market for standard transaction sizes.

B. T he exogenous liquidity risk corresponds to the normal variation of bid/ask spreads across
instruments.

C. T he endogenous risk of collective portfolio adjustments is easier to include in a VaR

computation.

D. T he exogenous component corresponds to the impact on prices of the liquidation of a

position in a relatively tighter market.

T he correct answer is B.

Exogenous liquidity risk can be, from a theoretical point of view, easily integrated into a VaR
framework. It corresponds to the normal variation of bid/ask spreads across instruments.
Liquidity risk can be divided into an endogenous liquidity component and an exogenous liquidity
component. Exogenous liquidity escapes from the control of the trader and finds its origin in the
characteristics of markets. T he key element of its determination is associated with precise modeling
of the spread’s behavior. By opposition, the endogenous liquidity component is specific to the
characteristics of the position and will, therefore, vary across markets.

100
© 2014-2023 AnalystPrep.
Q.1536 Expected shortfall is the most well-known risk measure following the VaR. It is conceptually
intuitive, has firm theoretical backgrounds and is now preferred to VaR. Which of the following
statements about the expected shortfall is/are correct?
I. T he expected shortfall does not account for the severity of losses beyond the confidence
threshold
II. T he expected shortfall is always sub-additive and coherent
III. T he expected shortfall mitigates the impact that the particular choice of a single confidence level
may have on risk management decisions

A. I and II

B. II and III

C. I and III

D. None of the above

T he correct answer is B.

T he expected shortfall does account for the severity of losses beyond the confidence threshold,
while always being subadditive and coherent, and mitigating the impact that the particular choice of a
single confidence level may have on risk management decisions.

Q.1537 In order to estimate the range and magnitude of the differences between compartmentalized
and unified risk measures, you are required to obtain a simple ratio of these two measures. After
performing relevant calculations, you come to the conclusion that ratio values greater than one
indicate:

A. risk compounding, and values less than one indicate risk diversification.

B. risk augmentation, and values less than one indicate risk reduction.

C. risk diversification, and values less than one indicate risk reduction.

D. risk augmentation, and values less than one indicate risk compounding.

T he correct answer is A.

Ratio values greater than one indicates risk compounding and values less than one indicate risk
diversification.

101
© 2014-2023 AnalystPrep.
Q.1538 T he examination of variable rate loans, in which the interaction between market and credit
risk can be analyzed, is performed by modeling the dependence of credit risk factors on the interest
rate environment. T hese factors include:
I. Loans' default probabilities
II. T he exposure at default
III. Loss-given-default

A. I and II

B. II and III

C. I and III

D. I, II and III

T he correct answer is D.

Credit risk factors include loans' default probabilities (PD), exposure at default (EAD), and loss-given-
default (LGD).

102
© 2014-2023 AnalystPrep.
Q.1539 Consider yourself as an intermediary who actively runs a VaR-based risk management system.
You start with a balance sheet consisting of risk-free debt and equity. T his is followed by an asset
boom, which leads to an expansion in the values of securities. Without any balance sheet adjustment,
this leads to:

A. A reduction in liability

B. An expansion in capital

C. An expansion in equity

D. A reduction in expenses

T he correct answer is C.

An asset boom leads to an expansion in the values of securities. Since debt was risk-free to begin

with, without any balance sheet adjustment, this leads to a pure expansion in equity.

Further Expl anati on

Value at risk - which is the equity capital the firm must hold to be solvent - is tied to a bank’s level of

economic capital. Most entities work with a target ratio of VaR to economic capital, and therefore a

VaR constraint on leveraged investors can be established. If there happens to be an economic boom,

it will relax the VaR constraint because the equity will now be worth more. T his will pave the way

for financial institution to take on more risk and further increase debt.

103
© 2014-2023 AnalystPrep.
Q.2647 Which of these is not a disadvantage of VaR?

A. VaR does not consider the worst case losses that lie beyond the VaR confidence level.

B. VaR is not subadditive.

C. T he VaR of different types of assets cannot be compared.

D. VaR gets difficult to calculate as the size of the portfolio and the number of assets in the
portfolio increases.

T he correct answer is C.

VaR is a simple and easy to understand and calculate metric for market risk. However, VaR is
calculated for a certain confidence level, and losses beyond that threshold are not considered.
Another disadvantage of the method is that the VaR measure is not subadditive. As the number of
assets in a portfolio increases, it gets more difficult to calculate VaR as the number of correlations
between the assets in the portfolio increase.

Q.2829 Let a balance sheet for an institution be given such that the liabilities side is 759 in debts and
104 equity shares. Calculate the total leverage.

A. 15.9

B. 21.5

C. 11.6

D. 8.3

T he correct answer is D.

Recall that leverage is the total assets to equity ratio. T his is given by the expression

Assets
L=
Equity

But from the accounting equation: Assets = Total Equities + Total Liabilities

⇒ Assets = 759 + 104 = 863

863
⇒L = = 8.3
104

104
© 2014-2023 AnalystPrep.
Q.2830 What type of liquidity risk is most troublesome for complex trading positions?

A. Endogenous

B. Spectral

C. Exogenous

D. Market-specific

T he correct answer is A.

Endogenous liquidity is the adjustment for the price effect associated with the liquidation of specific
positions. It is especially relevant in complex high-stress market conditions.

Q.2832 Branch Bank has the proportion of capital to be held per total VaR as 2.9 while the future
value of its assets is $56 million. Calculate the leverage for Branch Bank if the Value at Risk is $1.6
million.

A. 13.5

B. 22.1

C. 12.07

D. 15.6

T he correct answer is C.

We were given that Leverage:

A 1 A
L= = ( )×( )
K λ V aR

Where A is the future value of assets, and λ is the proportion of capital to be held per total VaR.

T hus:

1 56
⇒L = ( ×( )) = 12.07
2.9 1.6

105
© 2014-2023 AnalystPrep.
Q.3014 In many banks, aggregate risk is defined using a rollup or risk aggregation model; capital, as
well as capital allocation, is based on the aggregate risk model. Which of the following is least likely
correct regarding risk aggregation?

A. T he top-down risk aggregation model assumes that a bank’s portfolio can be cleanly
subdivided according to market, and operational risk measures only.

B. T he bottom-up risk aggregation model attempts to account for interactions among various
risk factors.

C. In the bottom-up aggregation model, the sub-risk levels are aggregated bottom-up using a
joint model of risk.

D. None of the above (All options are correct).

T he correct answer is A.

A is incorrect. In the top-down aggregation, risk is measured on the sub-risk level, and risks

considered are market risk, credit risk and operational risk.

B and C are both correct. In the bottom-up aggregation model, the sub-risk levels are aggregated

bottom-up using a joint model of risk and correlations between the different sub-risks that drive risk

factors.

106
© 2014-2023 AnalystPrep.
 Reading 69: Correlation Basics: Definitions, Applications, and
Terminology

Q.1521 Financial institutions determine the market values of forward contracts on the basis of some
underlying pricing factors. T hese factors can include market interest rate, risk, and correlation, etc.
Such factors can affect the price on the basis of their volatility and VaR percentage. What does the
positive correlation between two factors - A and B - indicate?

A. It indicates that when factor A goes up in value, the value of factor B is likely to
depreciate.

B. It indicates that when factor A goes down in value, the value of factor B is likely to
appreciate.

C. It indicates that when factor A goes up in value, the value of factor B does remains
unaffected.

D. It indicates that when factor A goes up in value, the value of factor B is likely to
appreciate.

T he correct answer is D.

A positive correlation means the prices of two assets move in the same direction. For example, a
correlation of 0.13 between the EURO spot and bill positions indicates that when the EUR goes up in
value against the dollar, the value of a 1-year EURO investment is likely to appreciate.

Q.1540 Financial correlation is the process of measuring the relationship between two or more
financial assets over time. It measures the extent to which two financial variables move with
respect to each other. T he original copula approach for collateralized debt obligation is a type of
static financial correlation that measures the default correlation of all assets in the CDO for a certain
time period. Here, the “certain time period” for a CDO is usually equal to:

A. the reinvestment period of the collateralized debt obligation.

B. the maturity of the collateralized debt obligation.

C. the time up to which the assets of the collateralized debt obligation defaults.

D. the time up to which any single asset of the collateralized debt obligation defaults.

T he correct answer is B.

T he “certain time period” refers to the time to maturity of a CDO.

107
© 2014-2023 AnalystPrep.
Q.1541 Assume that an investor has bought $2 million in a bond from Issuer A. T hey are now worried
about Issuer A defaulting and have purchased a Credit Default Swap (CDS) from Issuer B. T he value
of the CDS is mainly determined by the default probability of the reference entity Issuer A. If the
correlation between issuer A and B increases, what will be the impact on the price of the CDS?

A. T he price of the CDS will decrease because there is a greater chance of joint default.

B. T he price of the CDS will increase because there is a greater chance of joint default.

C. T here will be no impact on the price of the CDS because it is working as a separate entity.

D. It may increase or decrease depending on the market and economic conditions of the
country.

T he correct answer is A.

An increasing p means a higher probability of the reference asset and the counterparty defaulting

together. If the correlation between Issuer A and the CDS issuer increases, the present value of the

CDS for the investor will decrease and he will suffer a loss.

Further explanation:

If the default correlation between the CDS and the reference entity is positive, what this implies is

that the default probabilities of the protection seller (PS) and the reference entity (bond issuer) tend

to show co-movement; if the PD of the issuer increases, the PD of the PS also increases. What does

that mean? T he CDS is unl i k el y to serve i ts purpose of compensating the protection buyer in the

event of default on the part of the bond issuer because there's a good chance the protection seller

will not be in a position to make the agreed-upon payment.

T herefore, if the default correlation is positive, the protection seller will have a hard time

convincing the buyer to pay a higher premium. T he buyer will only accept a lower premium to

offset the risk of joint default.

108
© 2014-2023 AnalystPrep.
Q.1542 Suppose a correlation swap buyer pays a fixed correlation rate of 0.28 with a notional value
of $10 million for one year for a portfolio of three assets. T he following are the realized pairwise
correlations of the daily log returns at maturity for the three assets: ρ2,1 = 0.7 ρ3,1 = 0.2 ρ3,2 = 0.03
Assuming that for all pairs, i > j, the payoff for the correlation swap buyer is equal to:

A. $0.28 millin

B. $0.31 million

C. $0.3 million

D. $0.25 million

T he correct answer is C.

T he payoff for the correlation swap buyer is given by:

Payoff = Notional Amount × (ρrealized − ρfixed)

Where:

2
ρrealized = ∑ ρi. j
n2
− n i>j
2
= 2 ∑ ρi. j
3 − 3 i>j
1
= (0.7 + 0.2 + 0.03) = 0.31
3

T hus,

Payoff = 10, 000, 000(0.31 − 0.28) = $300, 000 = 0.3 milliom

109
© 2014-2023 AnalystPrep.
Q.1544 Nowadays in financial markets, investors are hedging their risk of portfolios by keenly
studying correlation and attempting to financially gain from those correlation changes. Correlation
trading is basically trading those assets whose prices are based on the movement of one or more
assets in time. In these correlation assets, the strike price - the price determined at the start of the
option - is commonly used. What does this strike price indicate?

A. T he price at which the underlying asset can be bought in the case of a call, and the price
at which the underlying asset can be sold in the case of a put.

B. T he price at which the underlying asset can be bought at the time the option is created.

C. T he price at which the underlying asset can be bought in the case of a put, and the price at
which the underlying asset can be sold in the case of a call.

D. T he right, but not the obligation, to buy or sell a stock at an agreed-upon price within a
certain period of time.

T he correct answer is A.

T he strike price refers to the price at which the underlying asset can be bought in the case of a call,
and the price at which the underlying asset can be sold in the case of a put. T he strike price is agreed
upon at the start of the option.
“T he right, but not the obligation, to buy or sell a stock at an agreed-upon price within a certain
period of time” is the definition of an option, not the strike price.

110
© 2014-2023 AnalystPrep.
Q.1545 When buying multi-asset options, investors must take into account any correlation between
the assets. In fact, the lower the correlation of two assets in an option, the higher the price of that
option. What does a negative correlation between the assets of the option indicate?
I. If one asset’s value decreases, on average, the other asset’s price appreciates.
II. If one asset’s value decreases, on average, the other asset’s price also decreases
III. If one asset’s value increases, on average, the other asset’s price appreciates.
IV. If one of the two assets appreciates, it will result in a high payoff, compensating the other asset’s
loss.

A. I only

B. I and III

C. I and IV

D. IV only

T he correct answer is C.

A negative correlation means prices move in different directions. If the price of one asset decreases,
on average, the price of the other one increases. T herefore, if one of the two assets appreciates, it
will result in a high payoff, compensating or reducing the other asset’s loss.

111
© 2014-2023 AnalystPrep.
Q.1546 A quanto option is another correlation option that authorizes a domestic investor to
interchange his potential option payoff (which is in foreign currency) back into its domestic
currency at a fixed exchange rate. T his option helps the investor get protected against currency
risk. From a risk management standpoint, the financial institutions which are selling these
correlation options do not have information about two things. T hese are:

A. the foreign currency amount to be converted into the domestic currency, and secondly,
the exchange rate at option maturity at which the foreign currency payoff will be converted
into the domestic currency.

B. the domestic currency amount to be converted into the foreign currency, and secondly,
the exchange rate at option maturity at which the domestic currency payoff will be
converted into the foreign currency.

C. the foreign currency that’s correlated with the domestic currency, and secondly, the
impact of the correlation on the buying and selling of quanto options.

D. the domestic currency amount to be converted into the foreign currency, and secondly,
the impact of the correlation on the buying and selling of quanto options.

T he correct answer is A.

T he financial institutions that sell quanto calls do not know two things:
1. How deep in the money the call will be, i.e., which Yen amount has to be converted into Dollars.
2. T he exchange rate at option maturity at which the stochastic Yen payoff will be converted into
Dollars.

112
© 2014-2023 AnalystPrep.
Q.1547 A correlation swap is a type of financial variable in which the correlation between assets can
be traded. In a correlation swap, a fixed known correlation is traded against the unknown correlation
that will actually occur. T his type of correlation swap protects the investor from a stock market
decline. T he payoff of a correlation swap for the correlation fixed rate payer at maturity is:

A. N (pfixed − prealised)

B. μ(prealised − pfixed)

C. N (prealised − pfixed)

D. N μ(prealised − pfixed)

T he correct answer is C.

T he payoff of a correlation swap for the correlation fixed rate payer at maturity is

N (ρrealised – ρfixed )

Where N is the notional amount.

113
© 2014-2023 AnalystPrep.
Q.1548 After the global crisis, financial institutions have become more risk-averse to avoid any
possible losses. For this reason, financial risk management has become a vital part of the financial
sector and VaR is one of the tools of financial risk management used to measure the market risk of
the portfolio. VaR measures the expected maximum loss of a portfolio with respect to a certain
probability for a time t. T he equation for VaR is:

VaR p = σp α √x

What do σp and α represent here?

A. σp is the volatility of the portfolio P, which includes the correlation between the assets in
the portfolio, while α is the abscise value of a standard normal distribution.

B. σp is the abscise value of a standard normal distribution, while α is the volatility of the
portfolio, which includes the correlation between the assets in the portfolio.

C. σp is the volatility of the portfolio P, which does not indicate anything about the
correlation between the assets in the portfolio, while α is the covariance matrix of the
returns of the assets.

D. σp is the volatility of the portfolio P, which includes the correlation between the assets in
the portfolio, while α is the covariance matrix of the returns of the assets.

T he correct answer is A.

α is the abscise value of a standard normal distribution corresponding to a certain confidence level
and σp is the volatility of the portfolio P , which includes the correlation between the assets in the
portfolio.

114
© 2014-2023 AnalystPrep.
Q.1549 We have calculated the value of VaR for a two-asset portfolio to analyze the impact of
correlations between the two assets. After going through all calculations of variance with the given
data, we reached a value of VaR. T he VaR value for a 10-day two-asset portfolio with a correlation
coefficient of 0.7 on a 99% confidence interval is $1.7486 million. What does this value imply?

A. Only once in a hundred 10-day period will this VaR amount ($1.7486 million) be exceeded.

B. We are 99% confident that we can lose more than $1.7486 million of our two asset
portfolio in the next 10 days.

C. We are 99% confident that we will not lose less than $1.7486 million of our two asset
portfolio in the next year.

D. Only once every 10,000 days will this VaR amount ($1.7486 million) be exceeded.

T he correct answer is A.

On average, only once in a hundred 10-day period (so once every 1,000 days) will this VaR amount
($1.7486 million) be exceeded.

Q.1550 Suppose we drew a graph showing the correlation between two assets, and found it to be
negative. It would mean that:

A. if the market value of one asset decreases, the other asset, on average, also decreases,
hence reducing the overall risk.

B. if the market value of one asset decreases, the other asset, on average, increases, hence
reducing the overall risk.

C. if one asset’s value decreases, the other asset’s value, on average, increases, hence
increasing the overall risk.

D. if one asset increases, the other asset, on average, increases, hence reducing the overall
risk.

T he correct answer is B.

T he lower the correlation, the lower the risk, as measured by the VaR. Preferably, the correlation is
negative. In this case, if one asset decreases, the other asset, on average, increases, hence reducing
the overall risk.

115
© 2014-2023 AnalystPrep.
Q.1551 T he global financial crisis of 2007-2009 was caused by a number of reasons which may
include high levels of debt, low interest rates, high-level speculation, and mortgage-backed securities.
It was the first correlation-related crisis marked by correlations among bonds and CDOs and this led
to the fall of many hedge funds.
Which statement is true regarding the cause of losses in CDOs?

A. T he losses occurred mainly from a lack of understanding of the correlation properties of

financial markets and hedge funds.

B. T he losses occurred mainly because of the correlation properties of the CDOs

themselves.

C. T he losses occurred mainly because of a lack of understanding of the correlation

properties of the tranches in the CDOs, not the CDOs themselves.

D. T he losses occurred because of an economic and financial downfall which eventually led
to the fall of the CDO market.

T he correct answer is C.

It is important to point out that the losses resulted from a lack of understanding of the correlation
properties of the tranches in the CDOs. T he CDOs themselves can hardly be blamed or be called
toxic for their correlation properties.

Q.1552 Other reasons for the financial crisis included residential mortgages, giving loans at lower
interest rates, and also the collapse of the subprime mortgage market. All these led to heavy selling
and buying of CDOs and Credit Default Swaps (CDSs).
. What is the purpose of the insurance contract underlying a Credit Default Swap?

A. To protect against the default of the underlying asset.

B. To speculate on the market movement in CDOs.

C. To reduce the individual risk of any asset by selling a large number of CDSs.

D. To trade a large number of securities at once to hedge against risk.

T he correct answer is A.

A CDC contract is similar to insurance in that it protects the contract buyer, that is, the owner of

the underlying credit, from default, or a credit rating downgrade.

In a CDC contract, credit risk is transferred between two parties.

116
© 2014-2023 AnalystPrep.
Q.1553 After the global financial crisis of 2007-2009, the financial regulators decided to implement
some rules and regulations for the stability of the financial markets and the banking sector. For this
purpose, the Basel Accords were introduced to deal with the deficiencies of the banking system. In
essence, the purpose of the Basel Accords was:

A. to offer motivations to the banking sector to improve their risk measurement and
management systems.

B. to offer motivations to investors to increase their investments in the banking sector.

C. to contribute to a higher level of trading in the banking system.

D. to invite risk managers and regulators to take part in the wellbeing of the financial sector.

T he correct answer is A.

T he objective of the Basel Accords was to “provide incentives for banks to enhance their risk
measurement and management systems” and “to contribute to a higher level of safety and soundness
in the banking system.”

Q.1554 Correlation risk is an important part of market risk which is typically measured with the help
of Value at Risk concepts. Market risk indirectly integrates the correlation risk. Market risk is also
measured using the expected shortfall, characterized as tail risk. Keeping this in mind, what is the
purpose of the expected shortfall?

A. To measure market risk for risky events, typically for the worst 0.1%, 1%, or 5% of past
scenarios.

B. It measures market risk for extreme events, typically for the worst 0.01%, 0.1%, or 1%
of possible future scenarios.

C. It measures market risk for extreme events, typically for the worst 0.1 %, 1%, or 5% of
possible future scenarios.

D. It measures market risk for risky events, typically for the worst 0.01%, 0.1%, or 1% of
past scenarios.

T he correct answer is C.

Market risk is also quantified with expected shortfall (ES), also termed conditional VaR or tail risk.
Expected shortfall measures market risk for extreme events, typically for the worst 0.1%, 1%, or
5% of possible future scenarios. Extreme events are those, despite having a very low probability of
occurrence, cause huge losses when they do occur.

117
© 2014-2023 AnalystPrep.
Q.1555 T he higher the default correlation between assets, the higher the probability that the
investors will lose all of their investments if a single asset’s price declines. However, lenders can
lower the default risk by diversifying their portfolio. What is the best policy for lending companies to
avoid risk and default?

A. Intersector default correlations are typically higher than intrasector default correlations
so the lending companies are recommended to create a sector-diversified loan portfolio to
decrease default correlation risk.

B. Intersector default correlations are typically lower than intrasector default correlations
so the lending companies are recommended to create a sector-diversified loan portfolio to
decrease default correlation risk.

C. Intrasector default correlations are typically higher than intersector default correlations
so the lending companies are recommended to specialize in a single-sector loan portfolio to
avoid default correlation risk.

D. Intersector default correlations are typically higher than intrasector default correlations
so the lending companies are recommended to specialize in a single-sector loan portfolio to
decrease default correlation risk.

T he correct answer is B.

Since the intrasector default correlations are higher than intersector default correlations, a lender is
advised to have a sector-diversified loan portfolio to reduce default correlation risk.

Q.1556 Systemic risk refers to the risks that affect financial markets as a whole. Which of the
following statements gives the correct relationship between systemic risk and correlation risk?

A. Systemic risk and correlation risk are partially dependent.

B. Systemic risk and correlation risk are highly independent.

C. Systemic risk and correlation risk are independent of each other.

D. Systemic risk and correlation risk are highly dependent.

T he correct answer is D.

Systemic risk and correlation risk are highly dependent. Since a systemic decline in stocks involves
almost the entire stock market, correlations between stocks increase sharply as a result of any
systemic decline.

118
© 2014-2023 AnalystPrep.
Q.1557 Concentration risk is a financial loss due to financial exposure. T his risk can be quantified
with the help of the concentration ratio. What will be the rule of thumb for the creditor regarding
the concentration ratio?

A. T he lower the value of the concentration ratio, the higher the diversification, and the
lower the default risk of the creditor.

B. T he higher the value of the concentration ratio, the higher the diversification, and the
lower the default risk of the creditor.

C. T he lower the concentration ratio, the lower the diversification, and the higher the
default risk of the creditor.

D. T he higher the value of the concentration ratio, the higher the diversification, and the
higher the default risk of the creditor.

T he correct answer is A.

T he lower the concentration ratio, the more diversified is the default risk of the creditor, assuming
the default correlation between the counterparties is smaller than 1.

Q.1558 Suppose that MLA Commercial Bank has lent Rs.10,000 to a single company, named X.
T herefore, MLA Commercial Bank’s concentration ratio is 1. In addition, suppose company X has a
default probability P(x) of 10%. T he expected loss from company X is Rs.10,000 x 0.1 = Rs.1,000.
Now, if MLA Commercial Bank lent the same amount of Rs.10,000 among three different companies,
assuming a default risk of 10% each, what will be the concentration ratio for the bank?

A. 0.5

B. 1

C. 0.3

D. 0.333

T he correct answer is D.

Concentration risk is the risk of financial loss due to a concentrated exposure to a particular group
of counterparties.
With a single borrower, concentration = 1. With three, the concentration ratio would be reduced to
1/3
Note:
Similarly, the concentration ratio for a creditor with 100 loans of equal size to different entities is
0.01 (= 1/100).

119
© 2014-2023 AnalystPrep.
Q.1559 Correlations and correlation risks form an important part of risk management because
different values of correlation result in different amounts of risk for any portfolio. Higher
correlations can lead to unexpected losses if not properly managed. T herefore, correlation trading
means:

A. traders should trade assets or implement trading strategies based on correlations between
assets.

B. traders should not trade assets or implement trading strategies on the basis of market and
economic conditions and how they related to single assets.

C. traders should trade assets or implement trading strategies on the basis of market and
economic conditions and how they related to single assets.

D. traders should trade assets or implement trading strategies based solely on uncorrelated
assets.

T he correct answer is A.

Correlation trading means that traders trade assets or execute trading strategies whose value is at
least in part determined by similar movements of two or more assets.

Q.1579 T he following are limitations of the Pearson correlation approach, EXCEPT :

A. T he Pearson correlation approach is typically not invariant to transformations. After

transformation of data, the information value of the Pearson correlation coefficient is
limited.

B. T he Pearson correlation approach only measures linear relationships and most financial
relationships are nonlinear

C. A zero correlation derived by the Pearson approach does not necessarily mean
independence. T herefore, the outcome of the Pearson correlation approach can be
misleading.

D. T he variances of the sets, say, X and Y, have to be infinite but finite for distributions with
strong kurtosis

T he correct answer is D.

T he variances of the sets X and Y have to be finite. However, for distributions with strong kurtosis,
for example, the Student's t distribution with v ≤ 2, the variance is infinite.

120
© 2014-2023 AnalystPrep.
Q.2634 A portfolio manager is considering adding one of two stocks to his existing portfolio. He has
gathered the following data to make his decision:

Annual Standard Value Correlation with

Deviation Portfolio
Existing Portfolio 20% $500
Stock 1 15% $100 0.5
Stock 2 25% $ 100 0.3

T he manager will only add a stock to the portfolio if the VAR of the resultant portfolio does not
exceed a daily VAR limit of $15 at a 99% confidence level. Given the information above, what should
the manager do?

A. Add Stock 1

B. Add Stock 2

C. Add either, if only VAR limit is the consideration, as the VAR of the resultant portfolio will
be the same in both cases

D. Add neither, as the VAR exceeds the VAR limit of $15 in both cases

T he correct answer is D.

Compute the variance of the portfolio:

(500 × 0.2)2 + (100 × 0.15)2 + 2(0.5)(500 × 0.2)(100 × 0.15) = 11, 725

VaR will be:

√11, 725 × 2.33

= 15.96
√250

Now compute the variance of the portfolio if Stock 2 is added:

(500 × 0.2)2 + (100 × 0.25)2 + 2(0.3)(500 × 0.2)(100 × 0.25) = 12, 125

VaR will be:

√12, 125 × 2.33

= 16.33
√250

T he daily VAR of both resultant portfolios exceeds the limit of $15, so the manager should invest in

121
© 2014-2023 AnalystPrep.
neither stock.

Additional explanation: Please note that the information given relates to a full year, particularly with

regard to volatility. However, the question requires us to compute the daily VaR. As such, we must

make use of the following relationship:

VaR(T days) = 1-day VaR × √T

T hus,

annual VaR
1-day VaR =
√T

We usually assume that there are 250 trading days in a year.

To compute the variance of a two-asset portfolio using the formula above, you could either use

weights or the actual market values of the respective positions:

(I) If you use weights, you'll end up with the percentage vol ati l i ty, i.e., the standard deviation of

the portfolio as a percentage of the total portfolio value. You'll then need to multiply the volatility by

the total size of the portfolio as a whole (of course as well as by the normal deviate -- 2.33 in this

case) to get the VaR in dollar terms.

(II) If you use the actual market values of each position, you'll end up with the dol l ar vol ati l i ty,

i.e, the standard deviation of the portfolio in dollar terms. In this case, you won't need to multiply the

volatility by the total size of the portfolio to get the VaR.

122
© 2014-2023 AnalystPrep.
Q.2639 For a portfolio having long positions in two assets, which value of correlation between the
assets will yield the highest value for VaR?

A. 1

B. 0.5

C. -0.5

D. 0

T he correct answer is A.

For a portfolio of two assets, if the assets are perfectly correlated, the value of VaR will be
maximized.

Q.2646 T he VaR of a portfolio is said to be undiversified when the assets in the portfolio are:

A. Positively correlated

B. Perfectly correlated

C. Negatively correlated

D. Not correlated

T he correct answer is B.

If the assets in a portfolio are perfectly correlated, then portfolio VaR is the sum of the individual
VaRs of the assets in the portfolio. In such a situation, the portfolio VaR is also known as
undiversified VaR.

123
© 2014-2023 AnalystPrep.
Q.2648 A portfolio consists of two assets X and Y. If $10 million is invested in the two assets in the
ratio 6:4 and the volatility of the two assets is 5% and 10% respectively, what will be the value of the
portfolio VaR at a 99% confidence level if the assets are (i) perfectly correlated and (ii)
uncorrelated?

A. Perfectly correlated: 1.16 million; Uncorrelated: 1.63 million

B. Perfectly correlated: 1.63 million; Uncorrelated: 1.16 million

C. Perfectly correlated: 1.35 million; Uncorrelated: 1.16 million

D. Perfectly correlated: 1.63 million; Uncorrelated: 1.35 million

T he correct answer is B.

Calculate the individual VaR of the two assets:

VaR X = P × (σz) = 0.6(10) × (0.05)(2.33) = 0.70 million

V aR Y = P × (σz) = 0.4(10) × (0.10)(2.33) = 0.93 million

When the assets are perfectly correlated, the portfolio VaR is simply the sum of the VaRs of the two

assets.

Portfolio VaR when perfectly correlated = 0.7 + 0.93 = 1.63 million

When the assets are uncorrelated, the portfolio VaR = √(V aR 2x + V aR 2y)

Portfolio VaR when uncorrelated = √(0.72 + 0.932) = 1.16 million

Q.2649 A portfolio is composed of two assets – A and B. An analyst has gathered the following daily
information about the portfolio:

Asset Value Return Standard Deviation

A 3 million 5% 3%
B 7 million 7% 5%

Assume that the correlation coefficient between asset A and asset B is 0.4. What will be the 1-day
VaR for this portfolio at a 99% confidence level under the variance-covariance approach?

A. 0.92 million

124
© 2014-2023 AnalystPrep.
 B. 0.85 million

C. 1.02 million

D. 0.36 million

T he correct answer is A.

T he covariances for the portfolio will be:

CovarianceAA = ρAA σAσA = 1 ∗ 0.03 ∗ 0.03 = 0.0009

CovarianceAB = ρAB σAσB = 0.4 ∗ 0.03 ∗ 0.05 = 0.0006

CovarianceBA = ρBA σBσA = 0.4 ∗ 0.05 ∗ 0.03 = 0.0006

CovarianceBB = ρBB σBσB = 1 ∗ 0.05 ∗ 0.05 = 0.0025

Cov(A, A) Cov(A, B)
[ ]
Cov(B, A) Cov(B, B)

T hus, our covariance matrix is:

0.0009 0.0006
[ ]
0.0006 0.0025

T he standard deviation for the portfolio can be calculated by multiplying the covariance matrix with

the amount of investment: We do this by first solving βh C , followed by βh C)βV , and then finding the

square root.

0.0009 0.0006
βh C = [ 3 7 ] [ ] = (3 ∗ 0.0009 + 7 ∗ 0.0006,3 ∗ 0.0006 + 7 ∗ 0.0025)
0.0006 0.0025
= [0.00690.0193]

3
βh CβV = [0.00690.0193] [ ]
7
= 0.009252 × 20 + 0.010384 × 10
= 3 × 0.0069 + 7 × 0.0193
= 0.1558

125
© 2014-2023 AnalystPrep.
T hus, σp = √0.1558 = 0.3947

Note: 0.3947 is the portfolio standard deviation in dol l ar terms,

VaR p = σp α√x = $0.3947 million × 2.33√1 = 0.92 million

Where: βh is the horizontal β vector of invested amounts

βY is the vertical β vector of invested amounts

C is the covariance matrix of the returns of the assets.

Note: When using the variance-covariance approach to calculate the VaR, the portfolio standard

deviation obtained is in dol l ar terms, not a percentage. In other words, variance-covariance VaR

uses dollar volatility (unlike parametric VaR which uses percentage volatility).

Note also, in this chapter, we will use a “variance-covariance VaR” approach to VaR calculations.

Q.2651 A bank has issued loans to two companies– A and B. T he probabilities of default of A and B are
5% and 10%, [Link] is the joint probability of default for the two companies if their
default correlation is 0.6?

A. 3.92%

B. 5.78%

C. 4.42%

D. 7.04%

T he correct answer is C.

P D = ρ(AB)√P D(A)(1 − P D(A)) × P D(B)(1 − P D(B)) + P D(A)P D(B)

Joint probability of default = 0.6√0.05(1 − 0.05) × 0.10(1 − 0.10) + 0.05 × 0.10

Joint probability of default = 0.6√0.004275 + 0.0050
Joint probability of default = 4.42%

126
© 2014-2023 AnalystPrep.
Q.2652 A bank has given companies A and B loans of $2 million each. Company A has a probability of
default of 5% while that of B is 15%. Calculate the expected loss of the bank under a worst-case
scenario if the default correlation is 0.5 and loss given default is 90%.

A. $167,040

B. $209,576

C. $172,312

D. $307,153

T he correct answer is A.

P D = ρ(AB)√P D(A)(1 − P D(A)) × P D(B)(1 − P D(B)) + P D(A)P D(B)

Joint probability of default = 0.5√0.05(1 − 0.05) × 0.15(1 − 0.15) + 0.05 × 0.15

Joint probability of default = 0.5√0.006056 + 0.0075
Joint probability of default = 4.64%

Expected loss = P D × LGD × EAD

Expected loss = 4.64% × 90% × 4, 000,000
Expected loss = $167, 040

Q.2653 What will be the most likely effect of a decreasing concentration ratio on the joint
probability of default given that the default correlation between the loans is less than 1?

A. It will increase

B. It is not possible to determine

C. It will decrease

D. T he concentration ratio does not affect the joint probability of default

T he correct answer is C.

T he concentration ratio will decrease (increase) as a bank issues more (less) loans. Unless the
default correlation between the loans is equal to 1, the decrease in concentration ratio will result in
a lower joint probability of default.

127
© 2014-2023 AnalystPrep.
Q.2834 Calculate the payoff of a correlation swap if the number of assets is 3 and the realized
pairwise correlations of the log-returns at maturity level are given as 0.24, 0.26, 0.36. You are also
given that the notional amount is $12 million at a 18% fixed rate with 1 year to maturity.

A. $0.29 million

B. $1.28 million

C. $0.24 million

D. $5.04 million

T he correct answer is B.

T he realized correlation is calculated as:

2
ρrealized = ( ) ∑ ρi,j
n2 − n i>j

T herefore,

2
ρrealized = ( ) (0.24 + 0.26 + 0.36) = 0.28667
2
3 −3

T hen:

N (ρrealized − ρfixed) = 0.28667 − 0.18 = 0.10667

T he payoff of the correlation fixed rate payer at maturity is

$12 million × 0.10667 = $1.28 million.

128
© 2014-2023 AnalystPrep.
 Reading 70: Empirical Properties of Correlation: How Do Correlations
Behave in the Real World?

Q.1560 Which of the following statements is/are INCORRECT about equity correlation?
I. Equity correlations fluctuate during expansionary and recessionary periods but in normal
economic periods, equity correlations do not fluctuate.
II. Economic stages only consider equity correlation volatility, not equity correlation levels.
III. T raders don’t need to consider higher equity correlation levels and higher equity correlation
volatility when making decisions.

A. I and II

B. I and III

C. II and III

D. I, II and III

T he correct answer is D.

All of the above statements about correlation are incorrect. Both equity correlation levels and
equity correlation volatility fluctuate in all the economic stages whether they are expansionary,
recessionary or normal. Moreover, traders and risk managers should take into consideration higher
correlation levels and higher correlation volatility that markets exhibit during times of economic
distress.

129
© 2014-2023 AnalystPrep.
Q.1561 T he equity correlation data of a particular country during different economic stages showed
that, in an expansionary economic stage, the correlation volatility was 74.54%. In normal economic
periods, the correlation volatility was 87.66%. In a recessionary economic stage, the correlation
volatility was 89.12%. Based on this information, we can conclude that:

A. correlation volatility is lower in recessions and normal economic periods but higher in
expansionary periods.

B. correlation volatility is higher in recessionary and normal economic periods but lower in
expansionary periods.

C. correlation volatility is higher in a recession but lower in normal and expansionary

periods.

D. correlation volatility is lower in a recession but higher in normal and expansionary

periods.

T he correct answer is B.

Equity correlation volatility is at it's lowest in an expansionary period and higher in normal and
recessionary economic periods.

Q.1562 An investor is willing to make an investment of $100 in a fixed-coupon bond. At maturity, this
bond will revert to exactly the par value of $100. T his type of bond gives an example of:

A. Autocorrelation

B. Mean reversion

C. Correlation volatility

D. Correlation levels

T he correct answer is B.

Fixed-coupon bonds, when they do not default, exhibit strong mean reversion: A bond is typically
issued at par, for example at $100. If the bond does not default, at maturity it will revert to exactly
that price of $100, which is typically close to its long-term mean.

130
© 2014-2023 AnalystPrep.
Q.1563 An investor is concerned about the figures of the mean reversion and autocorrelation of a
particular set of data. After analyzing the data, it is found that the mean reversion is 72.96%. Which
of the following is closest to the autocorrelation for the data?

A. 0.8171

B. 0.8935

C. 0.8296

D. 0.2704

T he correct answer is D.

Autocorrelation is the opposite property of mean reversion: the stronger the mean reversion, the
lower the autocorrelation.
T he sum of the mean reversion rate and the one-period autocorrelation rate will always equal one (=
72.96% + 27.04%).

Q.1565 T he economy of a country is experiencing a positive growth rate but, in the past, the
economy of the country faced recessionary stages several times. An investor is willing to make an
investment but he is not sure that the economy will remain in the expansionary stage in the near
future. How can future recessions be predicted in the country?

A. By using equity correlation, mean reversion and autocorrelation.

B. By using future inflation rates.

C. By the occurrence of a downturn in equity correlation volatility.

D. By analyzing growth rates.

T he correct answer is C.

Before every recession, a downturn in correlation volatility occurs. T herefore, it is likley that
equity correlation volatility is an indicator of a future recession.

131
© 2014-2023 AnalystPrep.
Q.1566 Bonds and their default probabilities also have correlation distributions just like equity. Which
of the following best describes the default probability correlation distribution and correlation
distribution for bonds?

A. Both the default probability correlation distribution and the correlation distribution can be
best modeled using the Johnson SB distribution.

B. T he default probability correlation distribution shows a normal shape and can be best
modeled using the generalized extreme value distribution, whereas the correlation
distribution can be best modeled using the Johnson SB distribution.

C. T he default probability correlation distribution can be best modeled using the Johnson SB
distribution, whereas the correlation distribution can be best modeled using the generalized
extreme value distribution.

D. Both the default probability correlation distribution and the correlation distribution can be
best modeled using the generalized extreme value distribution.

T he correct answer is C.

T he default probability correlation distribution is similar to the equity correlation distribution and
can be replicated best using the Johnson SB distribution. However, the bond correlation distribution
shows a more normal shape and can be best modeled using the generalized extreme value
distribution. It is to note that the bond correlation distribution can also quite well be replicated using
the normal distribution.

Q.2654 Given the following data about a variable S: St−1 = 40 St = 60 Mean reversion rate = 0.5
Calculate the long-run mean value for the variable.

A. 80

B. 60

C. 100

D. 75

T he correct answer is A.

St − St−1 = a(μ − St−1)

60 − 40 = 0.5(μ − 40)
20 = 0.5μ − 20
μ = 80

132
© 2014-2023 AnalystPrep.
Q.2655 Which of these distributions best fits an equity correlation distribution?

A. Chi squared

B. Generalized extreme value

C. Pareto

D. Johnson SB

T he correct answer is D.

T he Johnson SB is the best-fit distribution for equity correlation distributions.

133
© 2014-2023 AnalystPrep.
Q.2835 A correlation data has a long term mean of 52.6%. T he averaged correlation was again
observed as 31.26% in April 2011 for the 30 x 30 Dow correlation matrices. Given that the average
mean reversion was 82.1% from the regression function for 40 years, what is the expected
correlation one month later?

A. 23.56%

B. 32.21%

C. 48.78%

D. 38.23%

T he correct answer is C.

Using equation:

St − St−1 = aμs − aSt−1

T hen:

St = a(μs − st−1 ) + St−1

Where a = 82.1%, μs = 52.6%, St−1 = 31.26%

T herefore,

St = 0.821(0.526 − 0.3126) + 0.3126 = 0.4878 = 48.78%

T he mean reversion rate of 82.1% increases correlation of 31.26% in April 2011 to an expected

correlation of 48.78% in May 2011.

134
© 2014-2023 AnalystPrep.
Q.2837 T he Dow correlation matrices for a given data set today have an average correlation of
0.1645 with a long-term mean of 0.1972. Compute the expected correlation for exactly one year
from today if the average mean reversion is 0.7512.

A. 20.00%

B. 52.12%

C. 51.22%

D. 18.91%

T he correct answer is D.

Using equation:

St − St−1 = aμs − aSt−1

T hen:

St = a(μs − st−1 ) + St−1

Where a = 0.7512, μs = 0.1972, St−1 = 0.1645

T herefore,

St = 0.7512(0.1972 − 0.1645) + 0.1645 = 0.1891 = 18.91%

Q.2838 Which of the following is the most appropriate definition of autocorrelation?

A. T he tendency of a variable to be pulled back to its original mean.

B. T he degree to which a variable is correlated to its past values.

C. T he apparent relationship between the variables.

D. T he relationship between two variables keeping all other variables constant.

T he correct answer is B.

A variable is usually correlated to its past values up to a certain degree. T his degree of correlation is
called autocorrel ati on.

135
© 2014-2023 AnalystPrep.
Q.4872 In December 2020, a set of data is given such that the Dow correlation matrices have an
averaged correlation of 0.1925 with a long-term mean of 0.2723. Compute the expected correlation
for January 2021 if the average mean reversion is 0.6715.

A. 20.00%

B. 52.12%

C. 51.22%

D. 24.61%

T he correct answer is D.

Using the equation

St − St−1 = aμs − aSt−1

T hen:

St = a (μs − St−1) + St−1

Where a = 0.6715,μs = 0.2723, St−1= 0.1925

T herefore, St = 0.67150.2723– 0.1925 + 0.1925 = 0.2461 or 24.61% .

136
© 2014-2023 AnalystPrep.
 Reading 71: Financial Correlation Modeling—Bottom-Up Approaches

Q.1588 Copula functions, when described clearly, split down into multiple univariate distributions.
For instance:

C[G1(U 1 ),… … ., Gn (U n )] = Fn [F1−1 (G1(U 1 )),… ., Fn−1 (Gn (U n ));P F

In this illustration, Fi−1 describes:

A. the correlation structure of Fn .

B. the inverse of Fi.

C. the joint cumulative distribution function.

D. the marginal distribution.

T he correct answer is B.

In general, suppose Gi (ui) ∈ [0, 1] is a univariate, uniform distribution with ui = u1 ,… un and i ∈ N (i is

an element of set N). T hen, we define a copula function as follows:

C[G1(u1 ), … , Gn (un )] = Fn [F1−1(G1 (u1)), … , Fn−1 (Gn (un ));ρF ]

where:

Gi (ui) are the marginal distributions that have no well-known properties

Fn is the joint cumulative distribution function

Fi−1 is the inverse of Fi

ρF is the correlation structure of Fn

Put in words, the above equation reads:

Given the marginal distributions G1 (u1) to Gn (un ), there exists a copula function that allows the

mapping of the marginal distributions G1 (u1) to Gn (un ) via F −1 and the joining of the (abscise values)

Fi−1(Gi (ui)) to a single, n-variate function Fn [F1−1(G1 (u1)), … , Fn−1 (Gn (un ))] that has a correlation

structure of ρF .

137
© 2014-2023 AnalystPrep.
Q.1589 Because of appropriate and well-suited properties of the Gaussian copula, it is among the
most widely used copulas in finance.
When applying the n-variate case, which of the following statements is correct if Gx (U x) is uniform?

A. T he N −1(Gx (U x)) are univariate normal, and Mn is standard multivariate normal.

B. T he N −1(Gx (U x)) are multivariate normal, and Mn is univariate normal.

C. T he N −1(Gx (U x)) are standard normal, and Mn is standard multivariate normal.

D. T he N −1(Gx (U x)) are standard normal and Mn is standard normal.

T he correct answer is C.

If the Gx (U x)) is uniform, then the N −1(Gx (U x)) is standard normal and Mn is standard multivariate
normal. For proof, see Copula Method in Finance by Cherubini et. al., 2004.

Q.1590 T he following equation gives the Gaussian default time copula:

CG D [Qi = (t),. . ., Qn (t)] = Mn [N −1(Q1 (t)),. . ., N −1(Qn (t)); pm ]

It reveals that the term N −1 maps the cumulative default probabilities Q of asset i for time t, Qi (t) to
the univariate standard normal distribution, percentile to percentile. Keeping this in mind, which of
the following statements is correct?

A. T he 5th percentile of Qi (t) is plotted to the 5th percentile of the standard normal
distribution.

B. T he 4th percentile of Qi (t) is plotted to the 5th percentile of the standard normal
distribution.

C. T he 5th percentile of Qi (t) is plotted to the 3rd percentile of the standard normal
distribution.

D. T he 5th percentile of Qi (t) is plotted to the 10th percentile of the standard normal
distribution.

T he correct answer is A.

T he term N −1 maps the cumulative default probabilities Q of asset i for time t, Qi (t), to the
univariate standard normal distribution. So the 5th percentile of Qi (t) is mapped to the 5th percentile
of the standard normal distribution, the 10th percentile of Qi (t) is mapped to the 10th percentile of
the standard normal distribution, and so forth.

138
© 2014-2023 AnalystPrep.
Q.1591 Let’s assume we have only two assets - A and B. Suppose we feed our data to the Gaussian
default time copula function. How many correlation coefficients would we find?

A. One correlation coefficient.

B. One correlation matrix P m .

C. Multivariate correlation coefficients.

D. T wo correlation matrices.

T he correct answer is A.

Strictly speaking, only the bivariate Gaussian copula is a one-parameter copula, the parameter being
the copula correlation coefficient. A multivariate Gaussian copula may incorporate a correlation
matrix, containing various correlation coefficients.

Q.1592 Suppose we wish to analyze two companies, A and B, using the Gaussian default time copula.
After plotting the cumulative probabilities percentile to percentile to a standard normal distribution,
which of the equations below would we end up with?

A. Mi [N −1(QA)(t), N −1(Qi(t)); ρ]

B. Mi [N −1(QA)(t), N −1(QB(t)); ρ]

C. M5 [N −1(Q1)(t), N −1(QB(t)); ρ]

D. M2 [N −1(Q1)(t), N −1(Qi(t)); ρ]

T he correct answer is B.

We have only n = 2 companies A and B in our example. T herefore, the Gaussian default time copula
equation reduces to the form in choice B above. In addition, note that we would have only one
correlation coefficient (ρ), not a correlation matrix.

139
© 2014-2023 AnalystPrep.
Q.1593 To find the default time of an asset which is correlated to the default times of other assets
using the Gaussian default time copula, we would first need to:

A. derive the sample of normal standard distributions.

B. derive the sample of the correlation matrix.

C. derive the sample of Mn (. ) from the multivariate copula.

D. derive any of the components listed above.

T he correct answer is C.

To derive the default time t of asset i, T i, which is correlated to the default times of all other assets
i = 1...n, we would first derive a sample Mn (. ) from a multivariate copula, in the Gaussian case,
Mn (. ) ∈ [0, 1] .

Q.1594 When deriving the default time copula of an asset which is correlated to the default times of
other assets using the Gaussian default time copula, what is taken as the input from the n-variate
standard normal distribution Mn ?

A. T he N-variate matrix

B. T he average matrix

C. T he default time of assets

D. T he correlation matrix

T he correct answer is D.

When estimating the default time of asset i, T i, we include the default correlation with the other
assets in the portfolio. T he default correlation matrix is an input of the n-variate standard normal
distribution Mn .

140
© 2014-2023 AnalystPrep.
Q.1595 When flexible copula functions were introduced in the field of finance, they became very
popular immediately. But after some time they drastically lost their importance due to which of the
following unfavorable events/causes?

A. T hey found out not be helpful in solving complex problems

B. T hey are tough to apply to all statistical problems

C. T hey fell into disgrace when the global financial crisis hit in 2007

D. T hey found out not be helpful in the banking sector

T he correct answer is C.

When flexible copula functions were introduced to finance in 2000, they were enthusiastically
embraced. Investors believed that they could rely on the copulas to establish correlations among
multiple assets. However, copulas fell into disgrace when the global financial crisis hit in 2007, in
which correlations played a major part in the now infamous (disastrous) outcome.

Q.1596 Copula functions are introduced to simplify statistical problems. T hey enable the joining of
multiple univariate distributions to a single multivariate distribution. Which statement truly supports
the above expression of facts?

A. T hey transform an n-dimensional function into a unit-dimensional function.

B. T hey transform a one-dimensional function into an n-dimensional function.

C. T hey transform a one-dimensional function into a matrix.

D. T hey transform a matrix into an n-dimensional function.

T he correct answer is A.

A copula function transforms an n-dimensional function on the interval [0, 1] into a unit-dimensional

one:

C : [0, 1]n → [0, 1] .

141
© 2014-2023 AnalystPrep.
Q.1597 To derive the default time of a large number of assets during a simulation, the correlated
default time of multiple assets, the sample of Mn (. ) , is found through which of the following?

A. T he copula decomposition

B. T he normal decomposition

C. T he cumulative decomposition

D. T he Cholesky decomposition

T he correct answer is D.

To derive the default time T of asset i, T i , which is correlated to the default times of all other assets
i = 1,. . ., n, we first derive a sample Mn (. ) from a multivariate copula via the Cholesky
decomposition.

Q.1647 You acquire an asset that is negatively correlated with the economy. When investments are
negatively correlated we can use them in risk management for diversifying, or mitigating, the risk
exposures relevant to the portfolio. T he holdings which exist in that asset allow you to reduce your
exposure to the economy. T herefore, that asset is said to have:

A. Zero risk premium

B. Positive risk premium

C. Negative risk premium

D. None of the above

T he correct answer is C.

If an asset is negatively correlated with the economy, investors would expect a return that’s below
the risk-free rate of return, that is, a negative risk premium.

142
© 2014-2023 AnalystPrep.
Q.2656 A Gaussian copula maps the marginal distribution of each variable to which of the following
distributions?

A. Lognormal distribution

B. Poisson distribution

C. Standard normal distribution

D. Binomial distribution

T he correct answer is C.

A Gaussian copula maps the marginal distribution of each variable to the standard normal distribution.

Q.2843 A function X' is a 2-dimensional copula (2-copula). T he following are properties of the 2-
dimensional copula X'. Which one is the odd one out?

A. If ∝1 and ∝2 are subsets of set φ containing 0 and 1, then Domain X' = ∝1 × ∝2

B. X'(ε,1) = ε and X'(1, δ) = δ for all ε ∈ ∝1 and δ ∈ ∝2

C. X' is grounded and 2- increasing

D. X (ε,1) = δ and X'(δ,1) = ε for all ε ∈ ∝1 and δ ∈ ∝2

T he correct answer is B.

T his is odd because, X'(ε,1) = ε ⇒ T he variable reached its maximum point which is 1. T he variables
of this function should be equal to the value of the other attribute. Similarly, when δ reaches its
maximum value, implying 1, then the value of the function should be equal to the other attribute that
is δ. T his is correctly captured by option D.

143
© 2014-2023 AnalystPrep.
Q.2845 Define a Gaussian Copula.

A. T he default probabilities for an investment grade

B. T he one parameter copulas cited showing the overview of popular copula functions

C. A function allowing mapping of marginal distributions through another inverse function to

a single n-variate standard normal distribution with a correlation structure

D. None of the above

T he correct answer is C.

A Gaussian copula maps the marginal distribution of each variable to the standard normal distribution
which, by definition, has a mean of zero and a standard deviation of one. It allows preservation if
individual distribution functions while introducing a correlation function.

144
© 2014-2023 AnalystPrep.
 Reading 72: Empirical Approaches to Risk Metrics and Hedging

Q.1598 In any financial investment, you will find embedded financial risks. T here are different
methods to minimize such risks which are possible if you are able to locate and assess them. Some of
them include the use of hedging and risk metrics. Which of the following statements stands T RUE
regarding the aforementioned methods?
I. Risk metrics and hedging are mechanisms to provide a quantitative measure of the hidden financial
risks associated with a financial investment.
II. Risk metrics and hedging are mechanisms to provide a quantitative measure of only the
idiosyncratic financial risks associated with a financial investment.
III. Hedging and risk metrics reflect the interdependency of the rates and terms associated with a
financial investment.

A. Both I and II

B. Both II and III

C. Both I and III

D. None of the above

T he correct answer is C.

Hedging and risk metrics try to rein in all risks, particularly systematic ones that cannot be
eliminated via diversification. Multifactor metrics and hedges are implicit assumptions about how the
rates of different term structures change relative to one another.

145
© 2014-2023 AnalystPrep.
Q.1599 Inflation has an impact on the rate of returns associated with different financial instruments
provided and traded in the market. Which of the following statements is FALSE regarding T IPS
(T reasury Inflation Protected Securities)?
I. T IPS provide a relatively low rate of return to investors.
II. T IPS compensates for inflation by providing an inflation risk premium.
III. T IPS are traded at relatively high yields or low prices because their cash-flows aren’t inflation-
protected.

A. Both I and II

B. Both II and III

C. Both I and III

D. None of the above

T he correct answer is B.

T IPS make real or inflation-adjusted payments by regularly indexing their principal amount
outstanding for inflation. Real earnings are earnings that have taken into account the ‘eroded’
purchasing power of money via inflation.

Q.1600 T he nominal rate, a significant term in Finance, is often seen as the stated/advertised interest
rate on a loan, excluding charges, fees and/or interest compounding. What is the CORRECT definition
of the nominal rate?

A. T he nominal rate is the real rate plus the inflation rate.

B. T he nominal rate is the real rate minus the inflation rate.

C. T he nominal rate is the real rate plus the interest rate.

D. T he nominal rate is the real rate minus the interest rate.

T he correct answer is A.

T he nominal rate is applied to those securities which are inflation-protected. To offset the loss in the
purchasing power of money, an inflation risk premium is incorporated into the nominal rate.

146
© 2014-2023 AnalystPrep.
Q.1601 A 20-year semiannual coupon bond has a DV01 of 0.18125. An investor wishes to hedge his
position in this bond with another 10-year semiannual coupon bond whose DV01 is equal to 0.11369.
Calculate the hedge ratio:

A. 1.85

B. 1.59

C. 1.2

D. 1.5

T he correct answer is B.

A hedge ratio determines the amount of par of the hedge position that needs to be bought or sold for

every $1 par value of the original position. T he goal of hedging is to lock in the value of a position

even in the face of small changes in yield. T he hedge ratio is given by:

DV01 of initial position

HR =
DV01 of hedging tool
0.18125
=
0.11369
= 1.59

Interpretation: For every $100 par value of the 20-year bond, short $159 of par of the 10-year bond.

147
© 2014-2023 AnalystPrep.
Q.1602 Which of the following statements stands T RUE about the following equation of the Least-
Squares Regression Analysis?

ΔN = α + βΔy tR + ϵ t
t

I. Changes in the real-bond-yield are represented by ΔN and changes in the nominal-yield are
t
represented by Δy t .II. Changes in the nominal-yield are the independent variable creating changes in
R

the real-bond-yield, which is the dependent [Link]. T he slope and intercept for the formula are
to be assumed as the investor’s best [Link]. T he error term shows the real-bond-yield’s change
from the model’s predicted change on any specified day.

A. Both I and II

B. Both II and III

C. All of the above

D. None of the above

T he correct answer is D.

ΔN and Δy tR represent the changes in the yields of the nominal and real bonds respectively, and
t
changes in the real yield and the independent variable are used to predict changes in the nominal yield
and the dependent variable. T he intercept and the slope need to be estimated from the data whereas
the error term is the deviation of the nominal yield change on a particular day from the change
predicted by the model.

148
© 2014-2023 AnalystPrep.
Q.1603 T he following table represents the regression analysis of changes in the yield of the
T reasury Bond on changes in the yield of the T IPS (T reasury Inflation-Protected Securities) by
means of the following equation:

Δy tN = α + βΔy tR + ϵ t

No. of Observations 229

R-Squared 56.3%
Standard Error 3.82
Regression Coefficients Value Std. Error
Constant 0.0503 0.2529
Change in Real Yield 1.0189 0.0595

Which of the following statement gives an INCORRECT interpretation of the table above?

A. T he change in real yield of 1.0189 means that if the real yield increases by 1.0189 basis
points, the nominal yield decreases by 1.0189 bps over the sample period.

B. T he regression constant term is more or less equal to zero, meaning that the yield doesn’t
trend upwards or downwards when the comparable yield isn’t changing.

C. α and β as are normally distributed under the assumption of least squares and sufficiency
of data.

D. T he table shows R-squared (56.3%) representing the variance of nominal yield’s changes
that can be studied.

T he correct answer is A.

T he change in real yield reported in the table is 1.0189, which says that, over the sample period, the
nominal yield increases by 1.0189 basis points.

149
© 2014-2023 AnalystPrep.
Q.1604 Hedging is an investment strategy used to minimize the risk of adverse asset price
movements. T he hedge is normally carried out by taking an offsetting position in a related financial
instrument. When hedging two securities, one with the real rate of return and the other with the
nominal rate of return, it should be understood that:

A. the risk of the two securities can be measured accurately by DV01 alone.

B. the risk of the two securities cannot be measured accurately by DV01 alone.

C. the risk of the two securities can be measured accurately by PV01 alone.

D. None of the above.

T he correct answer is B.

T he risk of the two securities cannot be measured accurately by DV01 alone, in part because the
regression-based and DV01 hedges are certainly not always close in magnitude, even in other cases of
hedging T IPS versus nominal.

Q.1605 A regression framework is a statistical mechanism in the scope of finance, used widely to
determine the strengths/weaknesses of a relationship between two variables (one dependent and the
other independent). When used for hedging purposes, it provides many advantages. However, the
downside is that:

A. it is able to give an estimation of the hedged portfolio’s volatility.

B. a comparison can be made by the trader of the volatility with expected gain for his verdict
on the attractiveness of the risk-return structure.

C. the average change in the nominal yield for a given change in the real yield can be
estimated by the trader and he can then make adjustments to the DV01 hedge accordingly.

D. no complete control can be made on the dispersion of the change in the nominal yield in
relation to the change in the real yield.

T he correct answer is D.

With respect to improving the DV01 hedge, all the work is done based on an estimation rather than
on actual figures and there is not much the trader can do about the dispersion of the change in the
nominal yield for a given change in the real yield.

150
© 2014-2023 AnalystPrep.
Q.1606 T he hedge coefficient makes regression-based hedging difficult because, with the passage of
time, the regression coefficient estimated at one point in time tends to change. It can be handled by:
I. Estimating the coefficient over different time spans
II. Using available recent data since the more up-to-date the data is, the better will be the result.
III. Assuming the coefficient is always equal to 1.

A. Both I and II

B. Both I and III

C. All of the above

D. None of the above

T he correct answer is A.

Although chances are that the results achieved may vary greatly without assuming any one value for
the coefficient, blindly assuming a β of one (as in DV01 hedging) is generally not a superior approach.

Q.1607 Eric Rich, a trader, is making a relative value trade by selling a U.S. T reasury bond and
correspondingly purchasing U.S. T reasury T IPS. Guided by the current spread between the two
securities, Eric decides to short $100 million of the nominal bond and simultaneously purchases 76.2
million of T IPS. Soon afterward, Eric's position is disrupted by a change in the yield on T IPS in
relation to nominal bonds. After running a regression, he determines that the nominal yield has
changed by 1.03540 basis points per basis point in the real yield. By how much should Eric adjust the
hedge?

A. $2 million

B. $2.7 million

C. $79 million

D. $1.0354 million

T he correct answer is B.

T he trader would need to adjust the hedge as follows:

$76.2 million x 1.0354 = $78.9 million

T hus, the trader needs to purchase additional T IPS worth $2.7 million.

151
© 2014-2023 AnalystPrep.
Q.1608 T he Principal Component Analysis is a practical framework utilized in the scope of Risk
Analysis and Management in Finance, enabling traders to better estimate risk and return on their
securities. It’s particularly useful because:

A. T he sum of the variances of the first two PCs is usually quite close to the sum of
variances of all the rates.

B. T he sum of the variances of the first three PCs is usually quite close to the sum of
variances of all the rates.

C. T he sum of the variances of the last three PCs is usually quite close to the sum of
variances of all the rates.

D. None of the above.

T he correct answer is B.

PC Analysis provides the empirical structure on which basis one can simply describe the structure
and volatility of each of only three PCs.

Q.1609 Which of the following statements is FALSE regarding the Principal Component (PC)
Analysis:
I. T he PC Analysis provides a mechanism of empirical regularity for regression analysis.
II. T he sum of the PCs’ variances equals the sum of the individual rates’ variances capturing the
volatility of the set’s interest rates.
III. T he sum of the variances of the first two PCs is usually quite close to the sum of variances of all
the rates.

A. I only

B. I and II

C. III only

D. None of the above

T he correct answer is C.

Statement III is false. PCs of rates provide empirical regularity: the sum of the variances of the first
three PCs is usually quite close to the sum of variances of all the rates. Hence, rather than
describing movements in the term structure by describing the variance of each rate and all pairs of
correlation, one can simply describe the structure and volatility of each of only three PCs.
Statements I and II are correct. T he PC Analysis provides a mechanism of empirical regularity for
regression analysis. T he sum of the PCs’ variances equals the sum of the individual rates’ variances
capturing the volatility of the set’s interest rates.

152
© 2014-2023 AnalystPrep.
Q.1610 Principal Component (PC) Analysis is unique in scope and helps investors to achieve
maximum gain from trading of securities. Which of the following is correct with regard to PCs?

A. T he PCs are uncorrelated with each other while individual interest rates are highly
correlated.

B. T he PCs are correlated with each other while individual interest rates are highly
uncorrelated.

C. T he PCs are poorly correlated with each other while individual interest rates are highly
correlated.

D. T he PCs are poorly uncorrelated with each other while individual interest rates are highly
uncorrelated.

T he correct answer is A.

While changes in individual rates are, of course, highly correlated with each other, the PCs are
constructed so that they are uncorrelated.

Q.1611 Which of the following statements is true regarding principal components (PCs)?

A. Each PC is chosen to have the maximum possible variance given all earlier PCs.

B. Each PC is chosen to have the maximum possible variance given all later PCs.

C. Each PC is chosen to have the minimum possible variance given all earlier PCs.

D. Each PC is chosen to have the minimum possible variance given all later PCs.

T he correct answer is A.

T he first PC explains the largest fraction of the sum of the variances of the rates, the second PC
explains the next largest fraction, etc.

153
© 2014-2023 AnalystPrep.
Q.1612 T he financial markets comprise of different financial rates and the common classification is
in terms of time-frame, namely short-term and long-term financial rates. Which of the following
statements is T RUE regarding short-term and long-term rates associated with the financial markets?

A. Long-term rates are more volatile than short-term rates.

B. Short-term rates are more volatile than long-term rates.

C. Both long-term and short-term rates are equally volatile.

D. None of the above.

T he correct answer is B.

Changes in short-term rates are determined by current economic conditions, which are relatively
volatile, while longer-term rates are determined mostly by expectations of future economic
conditions, which are relatively less volatile.

Q.1614 What makes the Principal Component Analysis (PCA) highly practical and doable in the scope
of hedging and risk metrics?

A. T he PCA is very useful in the construction of empirically-based hedges for large

portfolios.

B. T he PCA is very useful in the construction of empirically-based hedges for small

portfolios.

C. T he PCA is very useful in the construction of theoretically-based hedges for large

portfolios.

D. T he PCA is very useful in the construction of theoretically based-hedges for small

portfolios.

T he correct answer is A.

It is impractical to perform and assess individual regressions for every security in a large portfolio.
In modern times, investors have large portfolios and through T he PCA, they can manage/hedge risks
quite seamlessly.

154
© 2014-2023 AnalystPrep.
Q.2846 Alvin Johnson is a trader and plans to short $230 million of the (nominal) 46⁄ 7s of 17th

February 2020 and purchase some amount of the T IPS 25⁄ 7 s of 17th January 2020 against that. T he

yields and DVO1s of a T IPS and a nominal US T reasury as of 30th April 2015 are provided as follows:

Bond Yi el d% DVO1

T ips 25⁄ 7 s of 17th January 2020 1.096 0.092

46⁄ 7 s of 17th February 2020 3.461 0.056

Compute the T IPS face amount that should be purchased for the trade to be hedged against the
interest rate levels.

A. $200 million

B. $275 million

C. $160 million

D. $140 million

T he correct answer is D.

T his can be achieved by making the trade DVO1-neutral.

Johnson has to buy the FR face amount of the T IPS such that:

FR × 0.092/100 = $230 million × (0.056/100)

⇒ FR = $230 million × (0.056/0.092)

= $140 million

Q.2847 A trader believes that the 5-year swap rate is far too high relative to the 2- and the 10-year
swap rates. She decides to receive in the 5-year and pay in the 10-year. T he tables below give the par
swap rates and the DVO1s of the swaps of the relevant terms. Compute the 5-year hedging ratio of
the DVO1, by the 2- and the 10-year swap.
Par swap rates and DVO1s:

155
© 2014-2023 AnalystPrep.
 Term Rate% DVO1

2 1.064 0.0286

5 2.535 0.0547

10 3.187 0.0931

Pri nci pal components of the USD swap curve:

Term Level Sl ope Short Rate PC Vol

2 5.27 -3.18 0.68 6.48

5 6.76 -1.64 -0.48 7.13

10 6.46 0.17 -0.45 6.27

Compute the 5-year hedging ratio of the DVO1, by the 2- and the 10-year swap.

A. 55.3% and 66.5%

B. 54.3% and 65.7%

C. 54.8% and 60.0%

D. 52.8% and 66.5%

T he correct answer is C.

T he equation that neutralizes the overall portfolio exposure to the level PC is:

0.0286 0.0931 0.0547

-F2 × 5.27 − F10 × 6.46 − 100 × × 6.76 = 0 .......equation 1
100 100 100

Similarly, the equation that neutralizes the overall exposure to the slope PC is:

0.0286 0.0931 0.0547

-F2 × (−3.18) − F10 × (0.17) − 100 × × (−1.64) = 0 ......equation 2
100 100 100

Solving for F2 and F10 simultaneously, we get -104.767 and -35.227 consecutively (T hese values are
calculated below)

In terms of the risk weights:

(104.767 × (0.0286/100))/0.0547= 54.8%

156
© 2014-2023 AnalystPrep.
And (35.227 × (0.0931/100))/0.0547= 60.0%

----------------------------------------
Sol vi ng si mul taneous equati ons:
T he two equations resulring from equations 1 and 2, respectively, are:
-0.150722F2 - 0.601426F10 = 36.9772 ....................equation 2
and
0.090948F2 - 0.015827F10 = -8.9708 ....................equation 3
From equation 3,
0.090948F2 = -8.9708 + 0.015827F10
T hus, F2 = -98.636584 + 0.174023F10
Substituting this in equation 2,
-0.150722[-98.636584 + 0.174023F10] - 0.601426F10 = 36.9772
14.866708 - 0.026229F10 - 0.601426F10= 36.9772
F10 = -35.227

Finally, F2 = -98.636584 + 0.174023F10 = -98.636584 + 0.174023(-35.227) = -104.767

157
© 2014-2023 AnalystPrep.
Q.2848 A trader plans to short $175 million of the 32⁄ 7s of 20th February 2019 and purchase some

amount of the T IPS 23⁄ 7 s of 20th January 2019 against that. Assume that the nominal yield in the data
changes by 1.036 basis points per basis-point change in the real yield. T he yields and DVO1s of a
T IPS and a nominal US T reasury as of 30th April 2015 are provided as follows:

Bond DVO1

23⁄ 7 s of 20th January 2019 0.087

32⁄ 7s of 20th February 2019 0.066

Compute the T IPS face amount that should be purchased for the trade to be hedged against the
interest rate levels.

A. $137.5 million

B. $126.2 million

C. $160.4 million

D. $140.0 million

T he correct answer is A.

T his can be achieved by making the trade DVO1-neutral.

T he trader has to buy the FR face amount of the T IPS such that:

FR × (0.087/100) = $175 million × (0.066/100) × 1.036

FR = $175mm × (0.066/0.087) × 1.036

= $137.5 million

158
© 2014-2023 AnalystPrep.
Q.2849 What is the role of regression analysis in bonds?

A. It is used to explain the value at risk of amounts invested in bonds

B. It explains the changes in the yield of one bond relative to the changes in yields of a small
number of other bonds

C. It is used for empirical description of the term structure of bonds

D. All the above

T he correct answer is B.

Regression analysis in bonds tries to explain the changes in the yield of one bond relative to changes
in the yields of a small number of other bonds.

159
© 2014-2023 AnalystPrep.
 Reading 73: The Science of Term Structure Models

Q.1616 T he market rate of a bond is considered to be equivalent to the price of that bond having the
same maturity. Securities with assumed prices are called underlying securities to distinguish them
from the:

A. proprietarily rights priced by arbitrage arguments.

B. contingent claims priced by arbitrage arguments.

C. contingent rights priced by arbitrage arguments.

D. proprietarily claims priced by arbitrage arguments.

T he correct answer is B.

Securities with assumed prices are called underlying securities to distinguish them from the
contingent claims priced by arbitrage arguments.

Q.1617 Refer to the following binomial tree:

1
p= ;5.50%
2

5%
↗
↘
1
p= ;4.50%
2

T he six-month rate is 5% today, which will be called date 0. On the next date six months from now,
which will be called date 1, there are two possible outcomes.
T he 5.50% state is called the:

A. Higher state

B. Upper extreme state

C. Up-state

D. Maximum state

T he correct answer is C.

In the figure above, the 5.50% state will be called the up-state while the 4.50% state will be called
the down-state.

160
© 2014-2023 AnalystPrep.
Q.1618 Refer to the following binomial tree:

1
p= ;5.50%
2

5%
↗
↘
1
p= ;4.50%
2

T he six-month rate is 5% today, which will be called date 0. On the next date six months from now,
which will be called date 1, there are two possible [Link] the current term structure of
spot rates, trees for the prices of six-month and one-year zero-coupon bonds can be computed. T he
price tree for $666 face value of the six-month zero will approximately be:

A. 635

B. 650

C. 675

D. 680

T he correct answer is B.

$666
= $649.75
0. 05
(1 + )
2

161
© 2014-2023 AnalystPrep.
Q.1620 We can price a security by means of arbitrage pricing and this is done by first searching and
valuing a portfolio which is a replica of the original one. In comparison, the derivative context is
complicated as its cash flows are dependent on the rates, and the portfolio replica is required to
duplicate the derivative security for any possible:

A. Inflation rate scenario

B. Call option scenario

C. Interest rate scenario

D. Sell option scenario

T he correct answer is C.

In the case of derivatives, cash flows depend on the levels of rates, and the replicating portfolio must
replicate the derivative security for any possible interest rate scenario.

Q.1621 An analyst wants to determine the value of a call option, maturing in six months, to purchase
$1,000 face value of a then six-month zero at $975. To do so, he constructs a replicating portfolio of
six-month and one-year zeros. T he following spot rates apply:
Six-month spot rate = 5.0%
One-year spot rate = 5.30%
Six months from now (date 1), the six-month rate will be either 6.0% (the up state) or 4.0% (the
down state). Both outcomes are equally likely to occur. Determine the value of the call option.

A. $2

B. $1.03

C. $0.5

D. $2.63

T he correct answer is B.

If the six-month rate on date 1 turns out to be 6.0% (the up state), the price of the then six-month
zero will be:
1000
1 =
970.87
(1+ 0. 06)
2
In this case, the right to buy the zero at $975 will be worth zero (T here’s no point exercising the
option).
If the six-month rate on date 1 turns out to be 4.0% (down state), the price of the then six-month
zero will be:
1000
1 =
980.4
(1+ 0. 04)
2
In this case, the right to buy the zero at $975 (call option strike) will be worth $5.4.

162
© 2014-2023 AnalystPrep.
To price the option by arbitrage, the analyst must construct a replicating portfolio today (date 0) of
underlying securities, namely six-month and one-year zero-coupon bonds, that will be worth $0 in the
up state on date 1 and $5 in the down state.
Let the face value of the six-month zero in the replicating portfolio be F0.5, and,
the face value of the 1-year zero in the replicating portfolio be F1 T hese values must satisfy the
following equations:
1
F0. 5 + F
1 1 ≡
F0. 5 + 0.9709F1 = $0.....in the up state
(1+ 0. 06)
2
1
F0. 5 + 1
F1 ≡ F0. 5 + 0.9804F1 = $5.4.....in the down state
(1+ 0. 04)
2
Solving these equations:
−0.9709F1 = 5.4 − 0.9804F1
F1 = 568.4211
F0. 5 = 5.4 − 0.9804 × 568.4211 = −551.8790
T hus, on date 0, the option can be replicated by buying about $568.4211 face value of one-year zeros
and simultaneously shorting about $551.8790 face amount of six-month zeros. By the law of one
price,
price of call option = price of the replicating portfolio

1 1
= F0. 5 + F1
0. 05 1 0. 053 2
(1 + ) (1 + )
2 2
1 1
=− × 551.8790 + × 568.4211 ≈ 1.033
1 2
(1 + 0.205) 0. 053
(1 + 2
)

Q.1622 You have been provided an interest rate tree to price the value of a one-year zero-coupon
bond and a call option on this bond. If the probability of an up-move increases suddenly, the current
value of a one-year zero should:

A. decline and the value of the call option should decline as well.

B. increase and the value of the call option should increase as well.

C. increase and the value of the call option should decrease.

D. decrease and the value of the call option should remain unchanged.

T he correct answer is A.

T he sudden increase in the probability of an “up-move” would translate into a higher likelihood that
interest rates will rise. As interest rates increase, bond prices decrease, and the value of call options
to purchase bonds must also fall. We can say that the price of an option is indirectly dependent on the
probabilities all the way through the price of the 1-year zero.

163
© 2014-2023 AnalystPrep.
Q.1624 T he risk penalty implicit in the call option price is inherited from the risk penalty of the one-
year zero, that is, from the premise that the price of the one-year zero is:

A. greater than its expected discounted value.

B. less than its expected discounted value.

C. equal to its expected discounted value.

D. not related to its expected discounted value.

T he correct answer is B.

T he risk penalty implicit in the call option price is inherited from the premise that the price of the

one-year zero is less than its expected discounted value.

Note that, price of the one year-zero is just the same as saying the expected value of the one-year

zero's price. Now, this is the value at time 1, before being discounted back to time 0. We now have

to discount this value at the given rate of interest and now what we get is the expected discounted

value of the one-year zero. T his value will obviously be less than the expected value of the one-year

zero at time 1.

Q.1625 It is a requirement of arbitrage pricing that the replica portfolio’s value must match the
option value in both ups and downs. One of the extraordinary aspects of this is that the real-world
probabilities of moving up and down are:

A. never encountered while calculating the arbitrage price.

B. always encountered while calculating the arbitrage price.

C. sometime encountered while calculating the arbitrage price.

D. partially encountered while calculating the arbitrage price.

T he correct answer is A.

T he real-world probabilities of up and down moves are never encountered i.e., they never enter into
the calculation of the arbitrage price.

164
© 2014-2023 AnalystPrep.
Q.1626 T he Black-Scholes-Merton model is not appropriate to value derivatives on fixed-income
securities because:

A. it assumes there is no upper limit to the price of the underlying asset.

B. it assumes bond price volatility is constant.

C. it assumes the risk-free rate is constant.

D. all of the above.

T he correct answer is D.

T he BSM model has several shortcomings which make it an inappropriate tool for valuing derivatives

on fixed-income securities:

(I) It assumes there is no upper limit to the price of the underlying asset. However, bond prices
actually have a maximum value.
(II) It assumes bond price volatility is constant. However, since bonds are redeemed at par, volatility
decreases as maturity approaches.
(III) It assumes the risk-free rate is constant. In reality, changes in short-term rates do occur,
causing rates along the yield curve and bond prices to change.

Q.1627 A call option, like a derivative, depends on the probabilities only through current bond prices.
If the probability of an up move suddenly increases, the current value of a one-year zero would
decline. If the replicating portfolio comprises of long one-year zeros, the value of the call option
would:

A. Increase

B. Decline

C. Remain unchanged

D. Cannot say due to insufficient information

T he correct answer is B.

If the current value of a one-year zero declines and the replicating portfolio is of long one-year
zeros, the value of the call option would decline as well.

165
© 2014-2023 AnalystPrep.
Q.1628 Risk-neutral pricing is a technique that modifies an assumed interest rate process so that any
contingent claim can be priced without having to construct and price its replicating portfolio. It is an
extremely efficient way to price many contingent claims under the same assumed rate process
because:

A. the original interest rate process has to be modified once or more than once, and this
modification only requires pricing the contingent claim(s) by arbitrage.

B. the original interest rate process has to be modified only once, and this modification only
requires pricing the contingent claim(s) by arbitrage.

C. the original interest rate process does not need to be modified, and there is no need to
price the contingent claim(s) by arbitrage.

D. None of the above.

T he correct answer is B.

In risk-neutral pricing, the original interest rate process has to be modified once, and this
modification only requires pricing a single contingent claim by arbitrage.

Q.1629 T he price of a derivative in the real economy may be computed as the discounted value
under the risk-neutral probabilities. Which of the following statement about the price of an option is
correct?

A. T he arbitrage price of the option equals its expected discounted value under the risk-
neutral probabilities.

B. T he price of a security that is priced by arbitrage depends on investors' risk preferences.

C. Investors in the imaginary economy penalize securities for risk and do not price securities
by expected discounted value.

D. T he price of an option does not necessarily need to be the same in the real and imaginary
economies.

T he correct answer is A.

T he arbitrage price of the option equals its expected discounted value under the risk-neutral
probabilities.

Q.1630 While performing arbitrage pricing in a multi-period setting, recombining trees are
considered to be economically reasonable. Consider the following tree diagrams and choose the

166
© 2014-2023 AnalystPrep.
correct option:
I.

6.00%

5.50%
↗
↘
5.00%
↗ 5.00%
↘
4.50%
↗
↘
4.00%

II.

6.00%

5.50%
↗
↘
5.00%
↗ 4.95%
↘
4.50%
↗
↘
4.05%

A. (I) is a recombining tree while (II) is a no recombining tree

B. (I) is a non-recombining tree while (II) is a recombining tree

C. Both (I) and (II) are recombining trees

D. Both (I) and (II) are non-recombining trees

T he correct answer is C.

A tree in which the up-down and down-up-states have the same value is called a recombining tree,

167
© 2014-2023 AnalystPrep.
whereas when an up move followed by a down move does not give the same rate as a down move
followed by an up move, the tree is said to be non-recombining.

Q.1631 You are asked to price a particular derivative security having a $200 million face value of a
stylized constant-maturity treasury swap struck at 5%. It is a one-year CMT swap on the six-month
yield. 50 bps increments/decrements are anticipated. On date 2, the state 2, 1, and 0 payoffs will be:

A. $1 million, -$1 million and $0 respectively

B. -$1 million, $0 and $1 million respectively

C. -$1 million, $1 million and $0 respectively

D. $1 million, $0 and -$1 million respectively

T he correct answer is D.

(y cmt − 5%)
Payoff = $200 million ∗ [ ]
2

Substituting 6%, 5% and 4% as Y CM T in will result in $1 million, $0 and -$1 million respectively.

Q.1632 An option-adjusted spread is a widely-used measure of the relative value of a security, that is,
of its market price relative to its model value. In addition, an option-adjusted spread can be elaborated
as spread which makes a security’s market price ______ the price of its corresponding model when
discounted values are computed at risk-neutral rates plus that spread.

A. equal to

B. greater than

C. lesser than

D. None of the above

T he correct answer is A.

An OAS is defined as the spread such that the market price of a security equals its model price when
the discounted value is computed at a risk-neutral rate plus that spread.

168
© 2014-2023 AnalystPrep.
Q.1633 T he return of a security or its profit and loss (P&L) may be divided into a component due to
the passage of time, a component due to changes in the factor, and a component due to the change in
the option-adjusted spread (OAS). On the other hand, if securities are non-priced in accordance with
the model (securities have an OAS greater/lesser than zero), their relevant cash flows are discounted
at:

A. T he short-term rate

B. T he long-term rate

C. T he short-term rate plus the OAS

D. T he short-term rate minus the OAS

T he correct answer is C.

Cash flows for securities that are not priced according to the model are discounted at the short-term
rate plus the OAS.

Q.1634 Usually, the time that elapses between dates of the tree is six months. However, we might
choose time steps smaller than six months because:
I. Decreasing the time step to a day, week, month or quarter assures that cash flows are adequately
close to pertinent data
II. Smaller steps result in a more realistic distribution of interest rates

A. I only

B. II only

C. All of the above

D. None of the above

T he correct answer is C.

In practice, we might choose time steps smaller than six months to bring cash flows closer to data
and attain a more realistic distribution of interest rates.

Q.2660 A constant maturity treasury (CMT ) swap of face value $1 million is struck at 6%. T he swap
(yCM T −6%)
pays 1, 000, 000 ( )where y CM T is a semiannually compounded yield, of a predetermined
2
maturity, on the payment date. Given the following binomial tree, calculate the value of the swap.

169
© 2014-2023 AnalystPrep.
 A. $678.22

B. $458.74

C. $798.12

D. $689.89

T he correct answer is B.

T he swap pays off when the interest rate exceeds 6%.

For the two scenarios that can occur in the first 6 months the payoff of the swap will be:

(6. 5%−6%)
Payoff if rate increases to 6.5% = 1, 000, 000 ( ) = $2, 500
2

(5. 5%−6%)
Payoff if rate decreases to 5.5% = 1, 000, 000 ( ) = −$2, 500
2

For the three scenarios that can occur after one year the payoff of the swap will be:

(7%−6%)
Payoff if rate increases to 7% = 1, 000, 000 ( ) = $5, 000
2

170
© 2014-2023 AnalystPrep.
 (6%−6%)
Payoff if rate remains at 6% = 1, 000, 000 ( ) = $0
2

(5%−6%)
Payoff if rate decreases to 5% = 1, 000, 000 ( ) = −$5, 000
2

(6%−6%)
Payoff if rate remains at 6% = 1, 000, 000 ( ) = $0
2

(5%−6%)
Payoff if rate decreases to 5% = 1, 000, 000 ( ) = −$5, 000
2

T he possible prices in six months are given by the expected discounted value of the 1-year payoffs

under the risk-neutral probabilities, plus the 6-month payoffs ($2, 500 and — $2, 500). Hence, the 6-

month values for the top and bottom node are as follows:

(5,000×0. 7+0×0. 3)
T he value of the upper node will be + 2, 500 = 5, 889.83
(1+0. 065)
2

(0∗0. 7−5,000∗0. 3)
T he value of the lower node will be − 2, 500 = −3, 959.85
(1+0. 055)
2

T he current price can then be calculated by multiplying the values calculated at 6 months by their

risk neutral probabilities and discounting them at the current rate:

(5, 889.83 ∗ 0.45 + (−3, 959.85 × 0.55))

Value = = $458.74
(1 + 0.206)

Q.2662 Which of the following is the term used to describe the process of valuing a bond using a
binomial interest rate tree?

A. Bootstrapping

B. Backward induction

C. Backtesting

D. Bootstrap historical simulation

T he correct answer is B.

T he process of valuing a bond using a binomial interest rate tree is known as backward induction.

171
© 2014-2023 AnalystPrep.
Q.2850 All of the following statements are true about callable bonds as screened against noncallable
bonds, EXCEPT :

A. At low yields, reinvestment risk falls

B. Capital gains are capped as the yield falls

C. T hey have less price volatility

D. T hey have negative convexity

T he correct answer is A.

Callable bonds have the following characteristics:

(I)When yields fall reinvestment risk i ncreases
(II)T hey have less price volatility
(III)T hey have negative convexity
(IV)Capital gains are capped as yields fall

172
© 2014-2023 AnalystPrep.
Q.2851 A $7 million face value of a stylized constant-maturity treasury (CMT ) swap is struck at 7%.
It is a one-year CMT swap on the six-month yield in 0.5% increments. Calculate the possible payoffs
of the CMT swap after 6 months and one year.

A. 6 months: $17,500 and -$17,500; One year: $35,000, 0 and -$35,000

B. 6 months: $35,000 and -$35,000; One year: $70,000, 0 and -$70,000

C. 6 months: $8,750 and -$8,750; One year: $17,000, 0 and -$17,000

D. None of the above

T he correct answer is A.

T he CMT swap pays:

y CMT − 7%
$7,000,000
2

In 6 months, the state 1 and 0 payoffs are, respectively:

7.5%−7%
$7,000,000 = $17,500
2

6.5%−7%
$7,000,000 = −$17,500
2

In one year, the state 2, 1 and 0 payoffs are:

8%−7%
$7,000,000 = $35,000
2

7%−7%
$7,000,000 = $0
2

6%−7%
$7,000,000 = −$35,000
2

173
© 2014-2023 AnalystPrep.
Q.2852 Which of the following statements are correct about the option-adjusted spread and the Z-
spread for option embedded bonds?

I. For a callable bond, the OAS is less than the Z-spread

II. For a putable bond, the OAS is greater than the Z-spread.

A. I only

B. II only

C. Both I and II

D. None of the above

T he correct answer is C.

With zero volatility, the Z-spread minus the OAS is equal to the option cost in percentage for a
callable bond greater than zero. For a putable bond, the OAS is greater than the Z-spread.

174
© 2014-2023 AnalystPrep.
 Reading 74: The Evolution of Short Rates and the Shape of the Term
Structure

Q.1636 You are asked to start with assumptions about the interest rate process and about the risk
premium demanded by the market for bearing interest rate risk and then derive the risk-neutral
process. T his approach results in:

A. An arbitrage-free model

B. An equilibrium model

C. T he matching of the initial term structure

D. T he pricing of all fixed-income securities by arbitrage

T he correct answer is B.

Models of this sort do not necessarily match the initial term structure and are called equilibrium
models.

Q.1637 One of the financial engineering models used in market analysis is known as the arbitrage-
free model. T his model basically allocates prices to instruments such as derivatives in a manner that
makes it extremely difficult to create arbitrage opportunities. In the case of an arbitrage-free model,
an understanding of the relationships between the model assumptions and the shape of the term
structure is important to:

A. make reasonable assumptions about the interest rate process and the risk premium.

B. comprehend the assumptions implied by the market through the observed term structure.

C. calculate the marginal productivity of capital.

D. compare propensity to save and expected inflation.

T he correct answer is B.

In the case of arbitrage-free models, an understanding of the relationships between the model
assumptions and the shape of the term structure reveals the assumptions implied by the market
through the observed term structure.

175
© 2014-2023 AnalystPrep.
Q.1638 T he 2-year spot rate, S2 is 9%, and the 1-year spot rate, S1 is 4%. What is the 1-year forward
rate?

A. 0.05

B. 0.048

C. 0.1024

D. 0.1424

T he correct answer is D.

Let the 1-year forward rate be y 1. T herefore:

(1 + s2)2 = (1 + s1)(1 + y 1)

(1.09)2
⇒ y1 = [ ] − 1 = 0.1424
1.04

Q.1639 As finance professionals, we know the definition of the term structure of interest rates as
“the relationship among bond yields or interest rates and different maturities or terms.” T he
forecasts are useful in describing the shape and level of the term structure over ______ horizons and
the level of rates at ______ horizons. T his carries significant implications when selecting term
structure models.

A. medium-term; short-term

B. long-term; short-term

C. long-term; medium-term

D. short-term; long-term

T he correct answer is D.

Forecasts are useful in describing the shape and level of the term structure over short-term
horizons and the level of rates in the long term.

176
© 2014-2023 AnalystPrep.
Q.1640 Convexity can make duration negative, since there are some securities, like few mortgage-
backed securities, that exhibit negative convexity. Assume that, other factors kept constant, the
value of convexity of the curve increases with maturity of its pricing function. T he securities with
greater convexity perform better when:

A. Yields remain constant

B. Yields change a little

C. Yields change a lot

D. None of the above

T he correct answer is C.

Securities with greater convexity perform better when yields change a lot and perform worse when
yields do not change by much.

Q.1641 Convexity is the rate at which the duration changes along the price-yield curve. In the case
of no interest rate volatility, the yields are completely determined by forecasts. But when volatility
is taken into account, the yields are affected by the value of convexity. T he value of convexity
increases with:
I. Volatility
II. Maturity
III. Yield

A. I only

B. I and II

C. I and III

D. I, II and III

T he correct answer is B.

T he value of convexity increases with volatility and maturity.

However, convexity typically decreases with an increase in yield. T he higher the interest rate (or

yield), the lower the convexity — or market risk — of a bond. T his reduction of risk occurs because

market rates would have to increase greatly to surpass the coupon on the bond, meaning there is less

risk to the investor.

177
© 2014-2023 AnalystPrep.
Q.1642 Assume that the maturity of a pricing function affects the convexity of the curve. A bond
with greater convexity is less affected by interest rates than a bond with less convexity. Also, bonds
with greater convexity will have a higher price than bonds with a lower convexity, regardless of
whether interest rates rise or fall. T herefore, which of the following securities perform worse
when yields do not change by much?

A. 1-year T-bill; convexity = 0.0265

B. 5-year T-note; convexity = 0.5863

C. 10-year T-note; convexity = 1.5986

D. 10-year corporate bond; convexity = 1.3256

T he correct answer is C.

Securities with greater convexity perform better when yields change a lot and perform worse when
yields do not change by much.

Q.1644 T he convexity in a financial model refers to non-linearities. For very short terms and
realistic levels of volatility, the value of convexity is quite small. It has been proved that convexity:

A. increases bond yields in theory and in practice.

B. decreases bond yields in theory and in practice.

C. increases bond yields in theory but not in practice.

D. decreases bond yields in theory but not in practice.

T he correct answer is B.

From market analysis, it has been found that convexity does, in fact, lower bond yields in practice.

178
© 2014-2023 AnalystPrep.
Q.1646 T he Capital Asset Pricing Model can be considered as the foundation of all financial domains
in this subject area but it also has prime relevance for practical decision-making. According to the
model, assets whose returns are positively correlated with aggregate consumption or wealth will
earn:

A. a return equivalent to the rate of the GDP.

B. a risk premium.

C. the risk-free rate of return.

D. a return equivalent to the rate of inflation.

T he correct answer is B.

T he asset pricing theory (e.g., CAPM) informs us that assets whose returns are positively correlated
with aggregate wealth or consumption will earn a risk premium.

Q.1649 While exploring the “relationship among bond yields or interest rates and different maturities
or term,” we can use also the term ‘yield curve’ to denote the term structure of interest rates. On
average, over the past 75 years, the term structure of interest rates has sloped upward. A term
structure that, on average, slopes upward can only be explained by:

A. a positive risk premium.

B. assumptions about convexity.

C. expectations of interest rate.

D. a negative risk premium.

T he correct answer is A.

Only a positive risk premium can explain a term structure that, on average, slopes upward.

179
© 2014-2023 AnalystPrep.
Q.2853 Investors value the current one-year interest rate at 11.316%. However, they also forecast
that for the following year, the one-year interest rate will be 13.457% and 15.658% for the year that
follows next. Calculate the two- and three-year spot rates, ρ(2) and ρ(3), respectively.

A. 11.3% and 13.5%

B. 15.3% and 16.1%

C. 12.8% and 13.5%

D. 12.4% and 13.5%

T he correct answer is D.

T he two-year spot rate ρ(2) is such that:

1 1
= =
(1.11316)(1.13457) (1 + p(2))2
1 1
⇒ =
(1 + p(2)) 2 (1.11316)(1.13457)
p(2) = √(1.11316 × 1.13457) − 1 = 0.12381 ≈ 12.4%

T he three-year spot rate is such that:

1
p(3) = (1.11316 ∗ 1.13457 ∗ 1.15658) 3 − 1
= 0.13463 ≈ 13.5%

180
© 2014-2023 AnalystPrep.
 Reading 75: The Art of Term Structure Models: Drift

Q.1650 We can determine the Continuously Compounded Interest Rate via the following simple
model when no drifting is considered and rates are normally distributed:

dr = σdw

In this equation, what do dr and dw indicate?

A. dr denotes the change in the time with the small change in interest rate measured
annually; dw indicates a normally distributed random variable

B. dr indicates the change in the rate over a small time interval, dt, measured in years; dw
indicates a normally distributed random variable with a mean of zero

C. dr indicates a normally distributed random variable with a mean of one; dw denotes the
change in the rate over a small time interval, dt, measured in years

D. dr denotes the change in the time with the small change in interest rate measured
annually; dw indicates a partially normal distributed random variable with a mean of one

T he correct answer is B.

dr indicates the change in the rate over a small time interval, dt, measured in years; dw indicates a
normally distributed random variable with a mean of zero.

181
© 2014-2023 AnalystPrep.
Q.1651 An FRM candidate is analyzing the movement in short-term interest rates assuming the rates
are normally distributed and there's no drift. Suppose the current short-term interest rate is 7.18%,
in a time interval of 3 months per year (or 3/12 per year), with a volatility of 118 basis points per
year. After a period of three months, the random variable dw has a value of 0.18. What are the
change in the short-term interest rate and the short-term rate after 3 months?

A. 0.2124% is the change in the short-term rate; 7.3924% is the short-term rate after 3
months.

B. 7.3924% is the change in the short-term rate; 0.2124% is the short-term rate after 3
months.

C. 1.2924 % is the change in the short-term rate; 8.4724% is the short-term rate after 3
months.

D. 8.4724% is the change in the short-term rate; 1.2924 % is the short-term rate after 3
months.

T he correct answer is A.

Using the formula for simple short-term rates, dr = σdw.

118
In this case, σ = and dw = 0.18
100

T hus,

118
dr = × 0.18 = 0.2124%
100

0.2124% is the change in the short-term rate; 7.3924% is the short-term rate after 3 months (7.18 +
0.2124).

182
© 2014-2023 AnalystPrep.
Q.1652 Term structure models in which the terminal distribution of interest rates has a normal
distribution are commonly known as Gaussian or normal models. T he limitation of these models is
that the short-term rate can become negative. Which of the following statements is true?

A. Negative short-term interest rate are very attractive for lenders because lenders will
have zero lending risk.

B. Individuals will never lend money at negative rates; they would rather hold it and earn a
zero rate.

C. Individuals will lend money at negative rates to help the borrowers in tough situations and
boost the economy.

D. Negative short-term rates can neither affect borrowers nor lenders because, in the long-
term, the rates become positive.

T he correct answer is B.

A negative short-term rate does not make much economical sense because people would never lend
money at a negative rate when they can at least hold cash and earn a zero rate instead.
Note the term "individuals" in the question. Some central banks or governments might lend a
negative rates in some special circumstances.

Q.1653 A popular method of overcoming the problem of negative interest rates is to construct
interest rate trees with the desired distribution and fix all negative rates to zero. When using this
method, rates in the original tree are considered as:

A. volatile market rates while the adjusted interest rates in the tree are called the interest
expected rates of interest.

B. volatile market rates of interest while the adjusted interest rates in the tree are called the
shadow rates of interest.

C. short-term rate of interest while adjusted interest rates in the tree are called the shadow
rates of interest.

D. shadow rates of interest while the adjusted interest rates in the tree are called the
observed rates of interest.

T he correct answer is D.

Rates in the original tree are called the shadow rates of interest while the rates in the adjusted tree
could be called the observed rates of interest. When the observed rate hits zero, it would “stay put”
until the shadow rate crosses back to a positive rate.

183
© 2014-2023 AnalystPrep.
Q.1654 T he simplest model of term structuring is Model 1. T his model’s term structure is downward
sloping, because it has no drifts and the rates decline uniquely with term. T he major aspect of this
model is the factor structure and the only factor of this model is the short-term rate. Now, suppose
this short-term rate increases by 20 basis points compounded semi-annually. What will be the impact
on the term structure?

A. Volatility will increase by 20 basis points because of an increase in the short-term rate.

B. Convexity will increase by 20 basis points because of an increase in the short-term rate.

C. Rates will increase by 20 basis points because of an increase in the short-term rate.

D. Maturity will increase by 20 basis points because of an increase in the short-term rate.

T he correct answer is C.

T he change in the term structure would be proportional. T herefore, all rates would increase by 20
basis points.

184
© 2014-2023 AnalystPrep.
Q.1655 Because of some limitations of Model 1(dr = σdw), another new term structure model was
introduced and named as Model 2. T he new model is written as:

dr = λ dt + s dw

Suppose r0 = 6.138% , λ = 0.239%, σ = 1.20% and the realization of the random variable dw happens
to be 0.15 over a month. Given these values, find the drift of the rate and standard deviation per
month, respectively.

A. T he drift of the rate is 0.239%, and the standard deviation is 0.18% per month.

B. T he drift of the rate is 0.239%; and standard deviation is 0.01992% per month.

C. T he drift of the rate is 0.01992%, and the standard deviation is 0.35% per month.

D. T he drift of the rate is 0.1992%, and the standard deviation is 0.18%% per month.

T he correct answer is C.

Note that the change in rate is given by the formula stated in the question, i.e., dr = λdt + s dw

However, we have been asked to find the drift of the rate and standard deviation per month,

respectively.

T he drift to the short-term rate is given by λdt while the standard deviation is given by σ√dt.
1 1
T herefore, λdt = 0.239% × 12 = 0.01992% and σdt = 1.20% √ 12 = 0.35%

185
© 2014-2023 AnalystPrep.
Q.1656 Model 1 and model 2 are usually known as equilibrium term structure models because of the
zero or constant drifts respectively. On the other hand, the model which varies with time is known
as the time-dependent drift. In this model, the drift depends on time and may vary from date to date.
From your understanding, what does the time-dependent drift represent over a period of time?

A. It represents some combination of the risk premium and some expected changes in the
short-term rate.

B. It represents some changes in the market short-term rates over time.

C. It represents expected changes in the volatility of short-term rates and market returns
over time.

D. It only represents changes in the risk premium occurring over a period of time.

T he correct answer is A.

T he drift that varies with time is called a time-dependent drift. Just as with constant drift, time-
dependent drift over each time period represents some combination of the risk premium and the
expected changes in the short-term rate.

Q.1657 Whenever any investor is buying or selling any financial instrument, it is of great importance
to match its price with changes in market prices. T he same case applies to the choice of the models
for term structure - whether to use arbitrage-free or equilibrium models. What is the most
important use of arbitrage-free models?

A. Quoting the prices of securities that are not actively traded based on the prices of more
liquid securities.

B. Quoting the prices of securities that are not actively traded based on time to maturity.

C. Quoting the prices of securities that are not actively traded on the basis of the economic
and financial stability of the lending institute.

D. Quoting the prices of securities that are not actively traded on the basis of prevailing
interest rate and swap rates.

T he correct answer is A.

One important use of arbitrage-free models is for quoting the prices of securities that are not
actively traded based on the prices of more liquid securities.

186
© 2014-2023 AnalystPrep.
Q.1658 Arbitrage-free models are used by practitioners for many purposes which include: valuing and
hedging many derivative securities, using real based assumptions to value the securities, etc. Why
are arbitrage-free models considered potentially superior to other security models?

A. T hese models are valuing the securities mainly based on time-dependent variables.

B. T hese models are valuing the securities mainly based on economic and financial reasoning.

C. T hese models are valuing the securities mainly based on the volatility assumptions and
sophisticated techniques.

D. T hese models are valuing the securities mainly based on parallel shift assumptions.

T he correct answer is B.

An arbitrage-free model is a financial engineering model that assigns prices to derivatives or other
instruments in such a way that it is impossible to construct arbitrages between two or more of those
prices. T he potential superiority of arbitrage-free models arises from their being based on economic
and financial reasoning.

Q.1659 A model matching market prices does not necessarily provide true values of the securities
and hedges for derivative securities. T he practice of fitting models to market prices is a good way to
incorporate the interest rate behaviors into the model but such a model may have some limitations
and warnings too. What are the main limitations of these types of models?
I. In some cases, adding a time-dependent drift to a parallel shift model to match a set of market
prices will make the model unsuitable for the intended application.
II. Expectation and risk premium built into the volatile assumptions of the model are not true
indicators of the security.
III. In many cases, market prices of the security or instrument are not fair in the context of that
model.
IV. T here are no limitations for these models because they incorporate market changes and prices.

A. I and II

B. I and III

C. IV only

D. II and III

T he correct answer is B.

Firstly, adding a time-dependent drift to a parallel shift model so as to match a set of market prices
will not make the model any more suitable for that application. Secondly, the argument for fitting
market prices assumes that those market prices are not fair in the context of the model.

187
© 2014-2023 AnalystPrep.
Q.1660 Mean reversion is the theory in finance which assumes that returns and prices will solely get
back to their mean or average values. T his mean (or average value) can be determined based on the
historical average or average returns and prices of that industrial sector. Assuming that the short-
term rate is characterized by mean reversion, what will be the effect on the rate if: (I) it is below
long-term equilibrium; and (II) if it is above long-term equilibrium?

A. (I) T he drift is positive, moving the rate up toward the long term value; (II) T he drift is
negative, moving the rate down towards the long-term value.

B. (I) T he drift is negative, moving the rate down towards the long term value; (II) T he drift
is positive, moving the rate up toward the long-term value.

C. T he drift is parallel based on parallel slope assumption and will behave irrespective of
changes in the long-term equilibrium.

D. Short term rates will change with the changes in the economic and financial condition of
that industrial sector irrespective of the changes in long-term values.

T he correct answer is A.

When the short-term rate is above its long-run equilibrium value, the drift is negative, driving the
rate down toward this long-run value. When the rate is below its equilibrium value, the drift is
positive, driving the rate up toward this value.

188
© 2014-2023 AnalystPrep.
Q.1661 T he risk-neutral dynamics of the Vasicek model can be written as:

dr = k(θ − r) dt + σdw

Here, the constant θ represents the long term value or the central propensity of the short-term rate
in the risk-neutral process, while “k” represents the quickness of mean reversion. What will happen
if the difference between r and θ increases?

A. T he greater the difference between r and θ, the greater the value of the short-term rate.

B. T he greater the difference between r and θ, the greater the value of k.

C. T he greater the difference between r and θ, the greater the expected change in the-short
term rate towards θ.

D. T he greater the difference between r and θ, the greater the expected change in the short-
term rate towards r.

T he correct answer is C.

In finance, the Vasicek model is a mathematical model describing the evolution of interest rates. It is
a type of one-factor short rate model as it describes interest rate movements as driven by only one
source of market risk. In such a scenario, the greater the difference between r and θ, the greater
the expected change in the short-term rate toward θ.

189
© 2014-2023 AnalystPrep.
Q.1662 Taking a numerical example which needs to be solved using the risk-neutral model of term
structuring, let k = 0.020,σ = 125 basis points per year, r∞ = 6.180% , λ = 0.227% , r = 5.212% .
Given these given values, find the expected change in the short-term rate and the standard deviation
over the next month?

A. A 36.08 basis point change in the short-term rate and a standard deviation of 1.70 basis
points.

B. A 1.70 basis points change in the short-term rate and a standard deviation of 36.08 basis
points.

C. A 2.053 basis points change in the short-term rate and a standard deviation of 36.08 basis
points.

D. A 35.04 basis point change in the short-term rate and a standard deviation of 1.54 basis
points.

T he correct answer is C.

T he Vasicek model equation appears as:

dr = k(θ − r)dt + σdw

λ
θ = r∞ + ( )
k
0.227
= 6.18 + ( ) = 17.53%
0.02

1
T he change in short-term rate = 0.020(17.53% − 5.212%) ( ) = 0.02053% = 2.053 basis points
12
1
T he volatility over the next month = 125 ∗ √( ) = 36.08 basis points
12

190
© 2014-2023 AnalystPrep.
Q.1663 Suppose we drew a graph showing the impact of the mean reversion on the terminal
distribution of short term rates. Risk-neutral distributions at different time horizons for the short-
term rate would show the impact of the mean reversion on the term structures. Which of the
following observations would we most likely make?

A. T he mean of the short-term rate, as a function of the time horizon, would remain constant
or relatively constant on the term structure.

B. T he mean of the short-term rate, as a function of the time horizon, would increase
gradually from its current value to its limiting value of θ.

C. T he mean of the short-term rate, as a function of the time horizon, will remain constant
from the current value to its limiting value of θ.

D. T he graph would show an increase in the mean of the short-term rate to match the
market volatility.

T he correct answer is B.

We would expect to see the mean of the short-term rate, as a function of the time horizon,

increasing gradually from its current value towards its limiting value of θ. In other words, the mean

of the short term rate is directly proportional to the horizon.

When the mean-reverting parameter k is relatively small, the mean of the short term rate rises very

slowly toward its limiting value of θ.

191
© 2014-2023 AnalystPrep.
Q.1664 T he mean-reverting parameter is not a particularly intuitive way of describing how long it
takes for a factor to return to its long-term goal. A more intuitive way is the half-life.
Suppose the half-life of interest rate is 28.72. What does this indicate?

A. T he interest rate factor takes 28.72 years to progress towards half of the distance
between its starting value and its goal.

B. T he interest rate factor takes half of 28.72 years to progress towards its goal.

C. T he interest rate factor, on a weighted average, takes 28.72 years to progress towards its
goal.

D. T he interest rate factor, on a weighted average, takes 57.44 years to progress towards its
goal.

T he correct answer is A.

A factor's half-life is defined as the time it takes the factor to progress half the distance toward its
goal.

192
© 2014-2023 AnalystPrep.
Q.1665 Using the Vasicek model, we can determine the standard deviation of the terminal distribution
of the short-term rate after T years.
Consider the following scenario:
A mean-reverting parameter has a value of 0.025 and volatility of 126 basis points. T he short rate in
10 years is normally distributed with an expected value of 7.4812%.

What is the standard deviation of the short rate in 10 years?

A. 343 basis points

B. 253 basis points

C. 243 basis points

D. 353 basis points

T he correct answer is D.

In the Vasicek model, the standard deviation of the terminal distribution of the short rate after T
years is given by:

σ2
= √[( )(1 − e−2kT )]
2k
1.262
= √[( )(1 − e−2×0. 025×10)]
2 × 0.025
= 3.53% or 353 basis points

193
© 2014-2023 AnalystPrep.
Q.1666 Financial institutions use mean-reverted as well as non-mean-reverted parameters to match
the par rates of securities with those of the market. Mean-reverted parameters can be used to fit
the model with observed term structures more accurately. After graphing the term structures of
mean-reverted as well as non-mean-reverted models, what would you expect?

A. T he model with mean reversion and the one without mean reversion would result in
dramatically different term structures of volatility.

B. T he model with mean reversion and the one without mean reversion would result in same
term structures of volatility.

C. Both models would be much more volatile and their patterns cannot be determined
through a small set of data.

D. T he model with mean reversion would give more accurate term structures than the one
without mean reversion.

T he correct answer is A.

A model with mean reversion and another one without mean reversion result in dramatically
different term structures of volatility. In a model with no mean reversion, rates are determined
exclusively by current economic conditions. Shocks to the short-term rate affect all rates equally,
giving rise to parallel shifts and a flat-term volatility structure. In a model with mean reversion,
short-term rates are determined significantly by current economic conditions, while longer-term
rates are determined significantly by long-term economic conditions.

194
© 2014-2023 AnalystPrep.
Q.1667 An FRM exam candidate draws a graph showing the volatilities of par rates with different
term structures including short-term as well as long-term term structures. Mean reversion and
volatility parameters are graphed against each other. T he model generates a term structure of
volatility that is sloping downwards, as mean reversion lowers the volatility of long term par rates.
From such a graph, we can conclude that:

A. the model matches the market at longer terms but understates the volatility for shorter
terms.

B. the model matches the market at shorter terms but overstates the volatility for longer
terms.

C. the model matches the market at longer terms but overstates the prices for shorter
terms.

D. the model matches the market at longer terms but overstates the volatility for shorter
terms.

T he correct answer is D.

Note that the volatilities of par rates decline with term in the Vasicek model. T he model in this case,
generates a term structure of volatility that is sloping downwards. When graph of volatilities against
different terms is plotted, the mean reversion and volatility parameters are chosen to fit the implied
two-point volatilities, say 2-and 10-year volatilities or longer terms, say 10-and-30 year volatilities. As
a result, it is easy to see that the model matches the market at those longer terms but overstates the
volatility for shorter terms. We can conclude that, Mean reversion lowers the volatility of longer
term par rates.

195
© 2014-2023 AnalystPrep.
Q.1668 An analyst draws a graph showing the effects of spot rates using the Vasicek model having a
10-basis-points change in the factor. T he graph interprets how the 10 basis point change in short-
term rate affects the spot rate curve. T he model is graphed with mean reversion. What would be the
effect on long-term and short-term rates given an increase of 10 basis points in the interest rate
factor?

A. T he short-term rates decrease by about 10 basis points but longer-term rates are less
impacted.

B. T he short-term rates increase by about 10 basis points but longer-term rates are less
impacted.

C. T he short-term rates increase by about 10 basis points but longer-term rates are impacted
by more than 10 basis points.

D. T he short-term rates, as well as long-term rates, increase by 10 basis points because of

the increase in volatility.

T he correct answer is B.

T he spot rate curve is affected by a 10 basis point increase in the short-term rate. By definition,
short-term rates rise by about 10 basis points but longer-term rates are impacted less. T he 30-year
spot rate, for example, rises by only 7 basis points. Hence, a model with mean reversion is not a
parallel shift model.

196
© 2014-2023 AnalystPrep.
Q.1669 Using whichever model for term structure, short-term rates are impacted by the changes in
economic and financial conditions of the markets. Long term-rates are less likely to be impacted by
shocks in the market but these incidents largely impact short term rates. Regardless of the changes,
the short-term change is assumed to arrive at long-term goals. Which of the following is correct
regarding short-lived and long-lived news?

A. T he news is short-lived if it changes the market's view of the economy many years in the
future; it is long-lived if it changes the market's view of the economy in the near future.

B. News is long-lived if it changes the market's view of the economy many years in the
future; it is short-lived if it changes the market's view of the economy in the near future.

C. Both types of news impact the economy in short term or long term irrespective of the
types of news.

D. Long-lived news impact long-term instruments and projects while short-lived news only
impact short-term instruments and securities.

T he correct answer is B.

Economic news is said to be long-lived if it changes the market's view of the economy many years in
the future. For example, news of a technological innovation that raises productivity would be a
relatively long-lived shock to the system. Economic news is said to be short-lived if it changes the
market's view of the economy in the near but not for future.

197
© 2014-2023 AnalystPrep.
Q.2661 Given the following data:
Current short-term interest rate: 1.5%
Long-run mean reverting level: 4%
Long-run true interest rate: 2%
Interest rate drift: 0.1%

Use the Vasicek model with mean reversion to determine the model’s mean-reverting parameter.

A. 0.040

B. 0.015

C. 0.022

D. 0.050

T he correct answer is D.

Mean reverting parameter:

λ
θ ≈η +
k

0.1%
4% ≈ 2% +
k

⇒ k = 0.05

Q.2663 Given the following binomial tree for the six-month spot rate:

198
© 2014-2023 AnalystPrep.
Calculate the expected discounted value of a one-year $1000 face value, zero-coupon bond.

A. $927.93

B. $971.82

C. $974.91

D. $951.13

T he correct answer is D.

199
© 2014-2023 AnalystPrep.
T he up and down values for the zero-coupon can be calculated by discounting the face value of the

bond by the up and down value for the six-month rate.

1000
Value if the six month rate is 6% = = 971.82
0. 058
(1+ )
2

1000
Value if the six month rate is 4% = = 978
0. 045
(1+ )
2

T he expected value of the one-year zero can be calculated by multiplying by the probabilities:

1 1
× 971.82 + × 978 = 974.91
2 2

T he discounted value of this can then be computed by discounting the expected value by the spot
974. 91
rate = 0. 05 = 951.13
(1+ )
2

200
© 2014-2023 AnalystPrep.
Q.2854 T he current value of a short-term rate is 7.31%. T he volatility is equivalent to 167 basis
points per year and the time interval under consideration is one month. Calculate the change in the
short-term rate in terms of basis points using the normally distributed rates and no drift model if after
a month, the random variable dw of mean zero and standard deviation of 0.2887 will take on the value
0.21.

A. 19 basis points

B. 35 basis points

C. 16.7 basis points

D. 731 basis points

T he correct answer is B.

Recall that: dr = σdw

From the information, r0 = 7.31%, σ = 1.67%, dw = 0.21

T herefore:
dr = 1.67% × 0.21
= 0.35% = 35 basis points

201
© 2014-2023 AnalystPrep.
Q.2855 Assume we are provided with a set of data whose current short-term rate is 6.047% with a
volatility of 207 basis points per year. We are also given a constant λ whose value is 0.348%. Using
the model of drift and risk premium, compute the change in rate if the monthly realization of the
random variable dw is 0.32.

A. 0.6914%

B. 0.0207%

C. 0.6047%

D. 3.4080%

T he correct answer is A.

From model 2 (drift and risk premium model), we have that:

dr = λdt + σdw

We are given that: λ = 0.348%, r0 = 6.047%, σ = 2.07% and dw = 0.32

and that the time interval under consideration is one month or 1⁄ 12 years.

Hence:
dr = 0.348% × 1⁄ 12 + 2.07% × 0.32
⇒ dr = 0.6914%

202
© 2014-2023 AnalystPrep.
Q.2856 We are given the following form of the Vasicek model:

dr = K(r∞ − R)dt + λdt + σdw = k([r∞ + λ/k] − r)dt + σdw

T he speed of mean reversion is 0.42 and the true interest rate process exhibits mean reversion to a
long-term value r∞ of 5.286%. Moreover, the current short-term rate is 4.256% and σ = 2.07%.
Using the model provided, determine the expected change in the short-term rate and the volatility
over the next month in basis points if λ = 0.256%.

A. 5.36 and 58.25 basis points

B. 5.75 and 55.32 basis points

C. 5.75 and 59.76 basis points

D. 65.51and 59.76 basis points

T he correct answer is C.

From the above information, we have that:

k = 0.42, r∞ = 5.286% , σ = 2.07%,r0 = 4.256% and λ = 0.256%.

T hus the expected change in the short rate is:

λ
dr = K(r∞ − R)dt + λdt + σdw = k ([r∞ + ] − r) dt + σdw
k

Note that:

θ ≡ r + λ/k

0.256%
∴ θ ≡ 5.286% + ( ) = 5.9%
0.42

1
dr = 0.42 × (5.9% − 4.256%) = 0.0005754 or 5.754 basis points
12

T he volatility over the next month is:

1
207 × √ = 59.76 basis points
12

203
© 2014-2023 AnalystPrep.
Q.2857 A dataset has a mean-reverting parameter of 0.37 and a volatility of 197 basis points. We are
also informed that the short-term rate is normally distributed. Use the Vasicek model to determine
the standard deviation of the terminal distribution of the short rate after 5 years.

A. 0.0336

B. 0.0556

C. 0.0226

D. 0.0198

T he correct answer is C.

Recall that the standard deviation of the terminal distribution of the short rate after T years is:

√σ2(1− e−2kT )/2k

T hus:

√0.01972(1− e−2×0.37×5)/(2×0.37) = 0.0226

204
© 2014-2023 AnalystPrep.
Q.4013 Under Model 1 of short-term interest rates, the current value of the short-term rate is
5.26%, that volatility equals 115 basis points per year, and that the time interval under consideration
1
is one month or 12 years. Mathematically, r0 = 5.26%; σ = 1.15%; and dt = 1/12. A month passes and
1
the random variable dw, with its zero mean and its standard deviation of √ 12 or 0.2887, happens to
take on a value of 0.25. Determine the short-term rate after one month.

A. 0.035

B. 0.025

C. 0.055

D. 0.002875

T he correct answer is C.

T he change in the short-term rate is given by:

dr = σdw
= 1.15% × 0.25 = 0.2875%

Since the short-term rate started at 5.26%, the short-term rate after a month is 5.55%.

New short-term rate = 5.26% + 0.2875% = 5.55%

205
© 2014-2023 AnalystPrep.
Q.4014 Under Model 2 of short-term interest rates, the current value of the short-term rate is
5.26%, that volatility equals 115 basis points per year, drift is 0.25%, and that the time interval under
1
consideration is one month or 12 years. Mathematically, r0 = 5.26%; σ = 1.15%; λ = 0.25%; and dt =
1 1
. A month passes and the random variable dw, with its zero mean and its standard deviation of √ 12 or
12
0.2887, happens to take on a value of 0.25. Determine the short-term rate after one month.

A. 5.5%

B. 5.5683%

C. 4.5212%

D. 0.3083%

T he correct answer is B.

Under Model 2,

dr = λdt + σdw

where:

dr = change in interest rates over small time interval, dt

λ = drift

dt = small time interval (measured in years) (e.g., one month = 1/12, 2 months = 2/12, and so forth)

σ = annual basis-point volatility of rate changes

dw = normally distributed random variable with mean 0 and standard deviation √dt

T hus, the change in the short-term rate is given by:

dr = λdt + σdw
1
= 0.25% × + 1.15% × 0.25
12
= 0.3083%

Since the short-term rate started at 5.26%, the short-term rate after a month is 5.5683%:

New short-term rate = 5.26% + 0.3083% = 5.5683%

206
© 2014-2023 AnalystPrep.
Q.4016 Using Model 1, assume the current short-term interest rate is 3%, annual volatility is 100bps,
and dw, a normally distributed random variable with mean 0 and standard deviation √dt, has an
expected value of zero. After one month, the realization of dw is -0.3. What is the change in the spot
rate and the new spot rate?

Change in spot New Spot Rate

I. 0.25% 2.25%
II. −3% 0.0%
III. 0.6% 3.6%
IV. −0.3% 2.7%

A. I

B. II

C. III

D. IV

T he correct answer is D.

T he change in the short-term spot rate is given by:

dr = σdw
= 1.0% × −0.3 = −0.3%

Since the short-term rate started at 3%, the short-term rate after a month is 2.7%.

New short-term rate = 3% − 0.3% = 2.7%

207
© 2014-2023 AnalystPrep.
Q.4017 T he U.S. department of T ransport has just announced an unexpected technological
breakthrough that will have a major bearing on the development of self-driving autonomous vehicles.
What is the most likely anticipated impact on a mean-reverting model of interest rates?

A. T he economic information is long-lived with a low mean-reversion parameter.

B. T he economic information is short-lived with a low mean-reversion parameter.

C. T he economic information is long-lived with a high mean-reversion parameter.

D. T he economic information is short-lived with a high mean-reversion parameter.

T he correct answer is A.

T he economic news is most likely long-term in nature. T herefore, the mean reversion parameter is
low so the mean reversion adjustment per period will be relatively low. Economic news is said to be
long-lived if it changes the market’s view of the economy many years in the future. T he
announcement of major technological progress in the development of self-driving cars would most
likely have a long-term and persistent effect on the automobile industry and the economy at large.

208
© 2014-2023 AnalystPrep.
Q.4018 Consider a Vasicek model with a reversion adjustment parameter of 0.03, annual standard
deviation of 200 basis points, a true long-term interest rate of 6%, a current interest rate of 5.0%,
and annual drift of 0.35%. Determine the expected rate in the model after 10 years:

A. 4.55%

B. 3.70%

C. 8.28%

D. 11.67%

T he correct answer is C.

T he expectation of the rate in the Vasicek model after T years is given by:

r0 e−kT + θ(1 − e−kT )

Where:

r0 = current interest rate

K = mean reversion parameter

θ = long-run value of the short-term rate assuming risk neutrality

T = time in years

First, we must work out the value of θ

λ 0.35%
θ ≈ rl + ≈ 6% + = 17.67%
k 0.03

rl is true long-term interest rate, and λ is the annual drift

T hus,

the expected rate after 10 years = 5%e−0. 03×10 + 17.67%(1 − e−0. 03×10 )
= 3.7041% + 4.5797%
= 8.2838%

209
© 2014-2023 AnalystPrep.
Q.4019 Consider a Vasicek model with a reversion adjustment parameter of 0.05, annual standard
deviation of 150 basis points, a true long-term interest rate of 6%, a current interest rate of 5.0%,
and annual drift of 0.4%. Determine the half-life of the model

A. 5.8

B. 26.0

C. 13.86

D. 14.50

T he correct answer is C.

Half-life is defined as the time it takes the factor to progress half the distance toward its goal. T he
half-life is given by:

ln(2)
Half-life = τ years =
k

Where k is the mean reversion parameter.

T hus,

ln(2)
half-life = = 13.86 years
0.05

210
© 2014-2023 AnalystPrep.
 Reading 76: The Art of Term Structure Models: Volatility and
Distribution

Q.1670 In the same way that we use a time-dependent drift to match the bond or swap rates, we can
also use time-dependent volatility functions to match option prices. T hese models focus on the
volatility of interest rates for term structure modeling. A simple time-dependent volatility function
can be written as:

dr = λ (t) dt + σ (t) dw

In this function, on which factor does the volatility of the short-rate depend?

A. T he volatility of the short-term rate depends on the interest rate.

B. T he volatility of the short-term rate depends on the volatility of σ.

C. T he volatility of the short-term rate depends on time.

D. T he volatility of the short-term rate depends on changes in the market prices of

securities.

T he correct answer is C.

T he volatility of the short-rate in the equation depends on time.

211
© 2014-2023 AnalystPrep.
Q.1671 Deterministic volatility functions and models are widely used by market makers to exploit
the benefits of interest rate options. Like in trading caplets, the value of caplets depends on the
distribution of short-rates at the time of expiration of these caplets. So, the flexibility of
deterministic functions can be used to match market prices of caplets with distinctive expiration
dates. At expiration, what does a caplet pay?

A. A caplet compensates the difference between the short rate and a strike, if positive.

B. A caplet compensates the difference between the short rate and a strike, if negative.

C. A caplet compensates the strike price at the time of expiration if the strike price
increases.

D. A caplet compensates the short rate only irrespective of the strike price at the time of
expiration.

T he correct answer is A.

At expiration, a caplet pays the difference between the short rate and a strike, if positive, on some
notional amount. Furthermore, the value of a caplet depends on the distribution of the short rate at
the caplet's expiration. T herefore, the flexibility of the deterministic functions may be used to
match the market prices of caplets expiring on many different dates.

212
© 2014-2023 AnalystPrep.
Q.1672 Model 3 is similar to the Vasicek Model with mean reversion in many ways. For example, if
the time-dependent drift of model 3 matches the average path of rates of the Vasicek model, then
both modes result in similar terminal distributions. However, these models differ in many ways.
Which of the following statements are true with regard to the differences between these two
models?
I. Model 3 is a parallel shift model just like models without mean reversion.
II. Model 3 is a parallel shift model just like models with mean reversion.
III. T he term structure of volatility is flat in Model 3, which is not the case with the Vasicek Model.
IV. T he term structure of volatility is curved in Model 3, which is not the case with the Vasicek
Model.

A. I and IV

B. II and III

C. I and III

D. II and IV

T he correct answer is C.

As is the case for any model without mean reversion, Model 3 is a parallel shift model. Also, the term
structure of volatility in Model 3 is flat. Since the volatility in Model 3 changes over time, the term
structure of volatility is flat at levels that change over time, but it is still always flat.

213
© 2014-2023 AnalystPrep.
Q.1673 Models with time-dependent volatility and those models with time-dependent drift with mean
reversion can be used for different securities based on the features of the securities. For example, if
you want to find the price of a fixed income option, then a model with:

A. time-dependent drift is suitable, but if you want to price and hedge fixed-income securities,
then a model with time-dependent volatility is preferable.

B. time-dependent volatility is suitable, but if you want to price and hedge fixed-income
securities, then a model with time-dependent drift is preferable.

C. time-dependent volatility is suitable, but if you want to price and hedge fixed-income
securities, then a model with mean reversion is preferable.

D. mean reversion is suitable, but if you want to price and hedge fixed-income securities,
then a model with time-dependent volatility is preferable.

T he correct answer is C.

If you want to quote fixed income options prices that are not easily observable, then a model with
time-dependent volatility provides a means of interpolating from known to unknown option prices. If,
however, the purpose of the model is to value and hedge fixed-income securities, including options,
then a model with mean reversion might be preferred.

214
© 2014-2023 AnalystPrep.
Q.1674 Many models of short-term rates assume the annualized standard deviation of dr is
independent of the interest rate level. T his makes the models irrelevant and inappropriate during
high inflation times and during periods of high-interest rates in the market. T herefore, a new CIR is
introduced:

dr = k(θ − r) dt + σ√rdw

Which of the following statements is correct regarding the CIR model above?

A. T he standard deviation of dr is inversely proportional to the square root of the short rate

B. T he standard deviation of dr is directly proportional to the square root of dw

C. T he standard deviation of dr is directly proportional to the square root of dt

D. T he standard deviation of dr is directly proportional to the square root of the short rate

T he correct answer is D.

T he annualized standard deviation of dr (i.e., the basis-point volatility) is proportional to the square
root of the rate. Put another way, in the CIR model, the parameter s is constant, but basis-point
volatility is not: annualized basis-point volatility equals σ√r and increases with the level of the short
rate.

215
© 2014-2023 AnalystPrep.
Q.1675 T he standard specification of the new CIR model is that the standard deviation of dr (basis-
point volatility) is proportional to the rate. In this model, σ is usually referred to as yield volatility,
and this specification leads to two different models: the Courtadon model and the Lognormal
[Link] the two models in mind, which of the following statements is correct regarding the
yield volatility and the basis-point volatility?

A. T he basis-point volatility is constant but the yield volatility equals σr and rises with the
level of the rate.

B. T he yield volatility is constant but the basis-point volatility equals σr and rises with the
level of the rate.

C. T he yield volatility as well as the basis-point volatility equal σr and both rise with the level
of the rate.

D. T he yeld volatility as well as the basis-point volatility equal σr and both decrease with the
level of the rate.

T he correct answer is B.

T he yield volatility is constant but the basis-point volatility equals σr and increases with the level of
the rate.

Q.1676 T he property of CIR model - that basis-points volatility equals zero when in situations when
short rate is zero - joined with the condition that drift is positive when the rate is zero, together
ensure that the short rate cannot move to negative values. In many aspects, this property of the CIR
model is an improvement over models with constant basis-point volatility.
Keeping this in mind, what is the problem of constant basis-point volatility with regards to interest
rates?

A. Models with constant basis-point volatility permit interest rates to become negative.

B. Models with constant basis-point volatility permit interest rates to become positive.

C. Models with constant basis-point volatility permit interest rates to change with market
changes in interest rates.

D. Models with constant basis-point volatility permit interest rates to change with spot rates.

T he correct answer is A.

T he CIR model is an improvement over models with constant basis-point volatility that allows
interest rates to become negative.

216
© 2014-2023 AnalystPrep.
Q.1677 T he choice of model mainly depends on the purpose at hand, that is, traders choose model for
term structure modeling on the basis of their purpose of buying or selling a security. Which of the
following statements are true if the traders opt for a constant volatility basis points model?
I. T he current economic environment is best with constant volatility.
II. T he current economic environment is very unstable and changes immediately with small changes
in investors’ perception.
III. T he negative rates have a minor impact on the valuing of the security under consideration.
IV. T he negative rates have a major impact on the valuing of the security under consideration.

A. I and IV

B. I and III

C. II and III

D. II and IV

T he correct answer is B.

A trader who believes that (I) the assumption of constant volatility is best in the current economic
environment, and (II) the possibility of negative rates has a small impact on the pricing of the
securities under consideration might very well opt for a model that allows some probability of
negative rates.

217
© 2014-2023 AnalystPrep.
Q.1678 A model was constructed to estimate the dynamics for a lognormal model with deterministic
drift. T he function of the model (time-dependent) is:

d[ln(r)] = α(t)dt + σdw

In this equation, the short rate has a lognormal [Link] this equation, what will be
the distribution of a random variable if its natural logarithm has a normal distribution?

A. T he random variable will be having a lognormal distribution if its natural logarithm has a
normal distribution.

B. T he random variable will be having a normal distribution if its natural logarithm has a
normal distribution.

C. T he random variable will be having a standard normal distribution if its natural logarithm
has a normal distribution.

D. T he random variable will be having an exponential distribution if its natural logarithm has a
normal distribution.

T he correct answer is A.

By definition, a random variable has a lognormal distribution if its natural logarithm has a normal
distribution.

Q.1679 A lognormal model with mean reversion is called the Black-Karasinski model. T his model
allows the volatility, mean reversion and short rate’s central tendency to depend on time. T hese
features make this model arbitrage-free. T his model shows that the natural logarithm of the short
rate is normally [Link] does this model allow the user to do which is not allowed in other
models?

A. A user can use or remove as much time dependence as desired.

B. A user can change the distributions as desired in any situation.

C. A user can change the price of the securities under consideration when needed.

D. A user does not have any extra privilege under this model.

T he correct answer is A.

A user may use or remove as much time dependence as desired under this time-dependent model.

218
© 2014-2023 AnalystPrep.
Q.1680 A lognormal model with mean reversion allows certain factors to depend on time, making it
an arbitrage-free model. T his model allows the user to make use of time dependence as desired for
the purpose at hand. T he dynamics of the model can be written as:

d[ln(r)] = k(t) [lnθ(t) − ln(r)] dt + σ(t)dw

T his equation assumes that the natural logarithm of short rates follows a time-dependent version of
the Vasicek model. Keeping this concept in mind, what is the distribution of natural logarithm short
rates in this equation?

A. T he lognormal distribution

B. T he normal distribution

C. T he standard normal distribution

D. T he Bernoulli distribution

T he correct answer is B.

According to the equation, the natural logarithm of the short rate is normally distributed. If viewed
another way, the natural logarithm of the short rate follows a time-dependent version of the Vasicek
model.

219
© 2014-2023 AnalystPrep.
Q.1681 T ime-dependent volatility functions or models are widely used by financial institutions
because of their flexible features. T hese models can be used to fit many option prices. A simple
volatility function suggests that the volatility of short rates depends on time. What does σ(1) = 1.25%
and σ(2) = 1.28% represent in that equation?

A. T he volatility of the short rate in 6 months is 125 basis points while the volatility of the
short rate one year is 128 basis points per year.

B. T he volatility of the short rate in one year is 128 basis points while the volatility of the
short rate in two years is 125 basis points per year.

C. T he volatility of the short rate in one year is 125 basis points while the volatility of the
short rate in two years is 128 basis points per year.

D. T he time-dependence of the short rate in one year is 125 basis points while the time-
dependence of the short rate in two years is 128 basis points per year.

T he correct answer is C.

If the function CY (t) were such that CY (1) = 1.25% and CY (2) = 1.28% , then the volatility of the
short rate in one year would be 125 basis points per year while the volatility of the short rate in two
years would be 128 basis points per year.

220
© 2014-2023 AnalystPrep.
Q.1682 A special case of the time-dependent volatility function is Model 3. Model 3 illustrates the
features of time-dependent volatility through the following equation:

dr = λ(t)dt + σe−αtdw

T his equation represents the behaviour of the short rate volatility. Which of the following
statements is true about the short rate volatility?

A. T he volatility of the short rate starts at the constant σ, and then exponentially decreases
to zero.

B. T he volatility of the short rate ends at the constant λ, and then exponentially decreases to
zero.

C. T he volatility of the short rate starts at the constant σ, and then exponentially increases
from zero to infinite.

D. T he volatility of the short rate starts at the constant λ, and then decreases to zero.

T he correct answer is A.

T he volatility of the short rate starts at the constant CY and then exponentially declines to zero in
this equation.

221
© 2014-2023 AnalystPrep.
Q.1683 T he choice of term structure depends on the purpose at hand. For instance, if the purpose of
the model is to price or hedge fixed-income securities/options, then the mean reversion model is
preferred because many users disagree with the time-dependent volatility model’s argument that
markets have a forecast of short term volatility in the distant future. Which modification in the new
model addresses this objection?

A. Assuming that volatility depends on time in the near future and then settles at a constant.

B. Assuming that the short rate depends on time in the near future and then settles at a
constant.

C. Assuming that the volatility depends on time in the distant future and then settles at an
increasing rate.

D. Assuming that volatility depends on time in the near future and then settles at a decreasing
rate.

T he correct answer is A.

T ime-dependent volatility relies on the difficult argument that the market has a forecast of short-
term volatility in the distant future. A modification of the model that addresses this objection is to
assume that volatility depends on time in the near future and then settles at a constant.

Q.1684 Which of the following statements is true about mean-reverting models?

A. T hey exhibit an upward-sloping factor structure and term structure of volatility which
capture the interest rate behavior movement much better compared to parallel shift models.

B. T hey exhibit a downward-sloping factor structure and term structure of volatility and are
as good at capturing interest rate behavior as parallel shift models.

C. T hey exhibit a downward-sloping factor structure and term structure of volatility and
capture interest rate behavior better than parallel shift models.

D. T hey exhibit upward-sloping factor structure and term structure of volatility and capture
interest rate behavior better than parallel shift models.

T he correct answer is C.

T he downward-sloping factor structure and the term structure of volatility in mean-reverting models
capture the behavior of interest rate movements better than models that yield parallel shifts or a flat
term structure of volatility.

222
© 2014-2023 AnalystPrep.
Q.1685 In the past, many models studied assumed that the basis-point volatility of the short rate was
independent of the level of the short rate, but in certain scenarios, this assumption went wrong,
making the models inappropriate for use (for instance, during times of high inflation). T his argument
led to a more specific model that considers basis point volatility of the short rate as an increasing
function.
Which of the following equations truly represents the dynamics of that model?

A. dr = k(λ − r)dt + σ√rdw

B. dr = k(θ − r)dt + σ√rdw

C. dr = k(θ − r)dt + σ√dw

D. dr = k(θ − λ)dt + σ√rdw

T he correct answer is B.

At extreme levels of the short rate, the basis-point volatility of the short rate is not independent of

the short rate. Also, in periods of high inflation, the short-term rate keeps fluctuating, which means

the basis-point volatility is quite high. Economic arguments of this sort have led to specifying the

basis-point volatility of the short rate as an increasing function of the short rate. T he risk-neutral

dynamics of the Cox-Ingersoll-Ross (CIR) model are:

dr = k(θ − r)dt + σ√rdw

223
© 2014-2023 AnalystPrep.
Q.1686 Models are always continually improved to make them more suitable in the face of changing
economic conditions. Many models with constant basis-point volatility allow interest rates to become
negative, which is not economically possible and appropriate.
Which property of model CIR guarantees that the short rate cannot become negative?

A. T he property that the basis-point volatility equals zero in case the short rate is zero,
joined with the condition that the drift is negative when the rate is zero, guarantees that the
short rate cannot become negative.

B. T he property that the basis-point volatility equals a constant in case the short rate is zero,
joined with the condition that the drift is positive when the rate is positive, guarantees that
the short rate cannot become negative.

C. T he property that the basis-point volatility equals a constant in case the short rate is zero,
joined with the condition that the drift is zero when the rate is zero, guarantees that the
short rate cannot become negative.

D. T he property that the basis-point volatility equals zero in case the short rate is zero,
joined with the condition that the drift is positive when the rate is zero, guarantees that the
short rate cannot become negative.

T he correct answer is D.

T he property that the basis-point volatility equals zero when the short rate is zero, combined with
the condition that the drift is positive when the rate is zero, guarantees that the short rate cannot
become negative. In some respects, this is an improvement over models with constant basis-point
volatility that allow interest rates to become negative.

224
© 2014-2023 AnalystPrep.
Q.2858 James Greenberg, an analyst at HSBC, is employing the Cox-Ingersoll-Ross (CIR) model for
the short-term rate process.
His assumptions include:

T he time-step is monthly, dt = 1/12, today's initial rate, r(0) = 2.11%, the annual basis point volatility,
sigma = 3.17%, the long-run rate, theta = 7.64%, the strength of reversion, k = 0.57.

For the first month, dw = 0.160. What is the short-rate in the first month under this CIR process,
r(1/12)?

A. -3.006%

B. -1.336%

C. 2.446%

D. 3.006%

T he correct answer is C.

Recall that from the CIR model, we have:

dr = k (θ − r)dt + σ√rdw
From the data provided in the question:
σ = 0.0317, r = 0.0211, θ = 0.0764, k = 0.57, dw = 0.16

T herefore:
dr = 0.57(0.0764 − 0.0211) × 1⁄ 12 + 0.0317√0.0211 × 0.16

⇒ dr = 0.00336 = 0.336%

T he short rate in the first month under this CIR model is:
2.11% + 0.336% = 2.446%

225
© 2014-2023 AnalystPrep.
Q.2859 What is the implication of basis-point volatility being equal to zero should the short rate be
zero in conjunction with the condition that a zero rate implies a positive drift?

A. T he long rate will always be positive

B. T he short rate will always be non-negative

C. T he long and the short rate will always be equal

D. None of the above

T he correct answer is B.

T he property that basis point volatility equals to zero when the short rate is zero combined with the
condition that the drift is positive when the rate is zero guarantees that the short rate cannot
become negative.

226
© 2014-2023 AnalystPrep.
 Reading 77: Volatility Smiles

Q.1705 T he Black-Scholes-Merton model is known in the field of Financial Risk Management for
depicting the variance of prices of instruments over a period of time. It is being used by traders
today but with a little variation from the method originally applied by Black, Scholes, and Merton and
this difference is because of:
I. T he allowance of the factor of stability to be used for pricing as an option to be dependent on its
strike-price
II. T he allowance of the factor of volatility to be used for pricing as an option to be dependent on its
strike-price
III. T he allowance of the factor of volatility to be used for pricing as an option to be dependent on its
time to maturity

A. Both I and II

B. Both II and III

C. Both I and III

D. None of the above

T he correct answer is B.

T raders do use the Black-Scholes-Merton model, but not in exactly the way that Black, Scholes, and
Merton originally intended. T his is because they allow the volatility used to price an option to depend
on its strike price and time to maturity.

Q.1706 T he term “volatility smile” carries significant value in the scope of Financial Risk
Management, and is used traders in equity and foreign currency markets. Which of the following
statements gives the correct definition of a volatility smile?

A. T he plot of volatility (implied) of financial markets determining prices of options.

B. T he plot of volatility (implied) of an option with a defined life span acting as a function of
its strike price.

C. T he plot of volatility (implied) of an option with an infinite life span acting as a function of
its stated price.

D. None of the above.

T he correct answer is B.

A plot of the implied volatility of an option with a certain life as a function of its strike price is
known as the volatility smile.

227
© 2014-2023 AnalystPrep.
Q.1707 T he term “volatility smile” is a pictorial representation of an option’s implied volatility and it
is being used by the traders as a pricing tool for their financial securities. Which of the following
statement is correct about volatility smiles?

A. T he volatility smile for European call-options with a certain maturity and strike price is
the same as that for European put-options with the same maturity and strike price.

B. T he volatility smile for European call-options with a certain maturity and strike price is
different from that for European put-options with the same maturity and strike price.

C. T he volatility smile for European call-options with a shorter maturity and same strike
price is the same as that for European put-options with a longer maturity and the same strike
price.

D. None of the above.

T he correct answer is A.

T he implied volatility of a European call option is the same as that of a European put option when
they have the same strike price and time to maturity. T he volatility smile provides their pictorial
representation.

228
© 2014-2023 AnalystPrep.
Q.1708 T he value of a foreign currency is $0.55. T he risk-free interest rate is 4% and 8% per year
in the U.S. and in the foreign country, respectively. T he market price of a European call option on
the foreign currency with a maturity of 1 year and a strike price of $0.57 is $0.0325. T he implied
volatility of the call is 14.5%. For there to be no arbitrage, the equation of the put-call parity
relationship is to be applied with q equal to the foreign risk-free rate. What is the value of a put
option according to the put-call parity?

A. 0.0725

B. 0.0724

C. 0.0419

D. 0.0687

T he correct answer is B.

Using the put-call parity:

p + S0e−qT = c + Ke−rT
p + 0.55e−0. 08∗1 = 0.0325 + 0.57e−0. 04∗1
p + 0.50771399 = 0.58014998
p = 0.58014998– 0.50771399
p = 0.72435989

229
© 2014-2023 AnalystPrep.
Q.1709 Which of the following statements stands T RUE for the following equation depicting the put-
call parity relationship in regard to the Black-Scholes-Merton model?

P BS − P m kt = CBS − Cm kt

I. BS in the equation reflects the inclusion of the Black-Scholes-Merton model

II. c and p represent the European call and put prices, respectively
III. T here is a no-arbitrage argument reflected in the equation
IV. Dollar error pricing is the difference between the value of a European option evaluated using the
Black-Scholes formula and the market value of the same option.

A. Both I and II

B. Both II and III

C. All of the above

D. None of the above

T he correct answer is C.

T he equation P BS − P m kt = CBS − Cm kt represents the put-call parity-relationship while using the

Black-Scholes-Merton model to price a European option with the same strike price and maturity time

leading to the same dollar pricing error.

Dollar error pricing is the difference between the value of a European option evaluated using the

Black-Scholes formula and the market value of the same option.

230
© 2014-2023 AnalystPrep.
Q.1710 After reading the following scenario, pick the statement which CORRECT LY depicts it.
Suppose that the implied volatility of a put option = 25% meaning that P BS = P m kt when volatility of
25% is being applied in the Black-Scholes-Merton model. From the following equation,
P BS − P m kt = CBS − Cm kt , then CBS = Cm kt when this volatility is used. T he implied volatility of the call
is also 25%.

A. T his scenario shows that the implied volatility of a European call option is always the
same as the implied volatility of a European put option when the two have the same strike
prices and maturity dates.

B. T his scenario shows that the implied volatility of a European call option is always the
different as the implied volatility of a European put option when the two have the same
strike prices and maturity dates.

C. T his scenario shows that the implied volatility of a European call option is always the
same as the implied volatility of a European put option when the two have different strike
prices and maturity dates.

D. T his scenario shows that the implied volatility of a European call option is always the
same as the implied volatility of an American put option when the two have the same strike
prices and maturity dates.

T he correct answer is A.

For a given strike price and maturity, the correct volatility to use in conjunction with the Black-
Scholes-Merton model to price a European call should always be the same as that used to price a
European put. T his means that the volatility smile is the same for European calls and European puts.

231
© 2014-2023 AnalystPrep.
Q.1711 When it comes to making a decision regarding trading using the Black-Scholes-Merton model,
assumptions must be made regarding the distribution of exchange rates. But at the same time, the
lognormal assumption is deemed as not a right choice for exchange rates, and it’s advisable to buy
deep-out-of-the money call and put options on a variety of different currencies and wait. T his
suggestion is supported by the following reasons:
I. T he chosen options will be relatively inexpensive
II. Many of the chosen options will close in the money than the prediction of lognormal-model
III. On average, the payoffs’ present value will be more than the options’ cost

A. Both I and II

B. Both I and III

C. All of the above

D. None of the above

T he correct answer is C.

From around the mid-1980s, a few traders knew about the heavy tails of foreign exchange probability
distributions. Everyone else thought that the lognormal assumption of Black-Scholes-Merton was
reasonable. T he few traders who were well informed followed the strategy that has been discussed
in the question made tons of money. By the late 1980s, everyone realized that foreign currency
options should be priced with a volatility smile and the trading opportunity disappeared.

Q.1713 In practice, exchange rates do not work on the condition of lognormal distribution as the
exchange rate’s volatility is far from constant, and there are frequent jumps. Which of the following
statements stands FALSE for this scenario?

A. Extreme outcomes are expected as a result of the impact of the variable volatility and
jumps.

B. T he effect of the variable volatility and jumps are independent of the maturity of options.

C. When the maturity of options increases, the volatility smile becomes less pronounced.

D. None of the above

T he correct answer is B.

As the maturity of the option increases, the percentage impact of non-constant volatility on prices
becomes more pronounced, but its percentage impact on implied volatility usually becomes less
pronounced. T he percentage impact of jumps on both prices and the implied volatility becomes less
pronounced as the maturity of the option is increased. T he result of all this is that the volatility smile
becomes less pronounced as option maturity increases.

232
© 2014-2023 AnalystPrep.
Q.1715 Which of the following holds true for equity options smirk?
I. Leverage is identified as one of the main reasons for the smirk in equity options
II. It is said that when a company's equity declines in terms of value, the company's leverage then
increases making equity riskier, and its volatility increases
III. It is said that when a company's equity increases in terms of value, the company's leverage then
decreases making equity less risky, and its volatility decreases

A. Both I and II

B. Both I and III

C. All of the above

D. None of the above

T he correct answer is C.

T he equity options volatility pattern is different from the currency option smile. T he pattern is

more of a “smirk,” than a smile.

T he company’s leverage and equity value are indirectly correlated, which has an impact on the

equity’s risk and volatility.

An increase in a firm’s equity results in a decrease in leverage, which tends to decrease the riskiness

of the firm. T his lowers the volatility of the underlying asset. On the other hand, a decrease in the

firm’s equity results in an increase in leverage, which makes the firm riskier. T his increases the

volatility of the underlying asset.

233
© 2014-2023 AnalystPrep.
Q.1716 T raders use volatility smiles to allow for non-log-normality when trading call and put options.
Which of the following statements stands T RUE about volatility smiles for equity options?
I. A volatility smile depicts the relationship between the option’s implied volatility and the option’s
strike price
II. Volatility smiles for equity options are drawn as downward slopes on graphs
III. Volatility smiles show that out-of-the-money calls tend to have higher implied volatility as
compared to in-the-money calls
IV. Volatility smiles show that in-the-money puts tend to have higher implied volatility as compared to
out-of-the-money puts

A. I and II

B. III and IV

C. All of the above

D. None of the above

T he correct answer is A.

Statement I is correct. A volatility smile defines the relationship between the implied volatility of an
option and its strike price.
Statement II is correct. For equity options, the volatility smile tends to be downward sloping. (Refer
to the image at the bottom of the page.)

Statement III and IV are incorrect. In-the-money calls and out-of-the-money puts tend to have high
implied volatilities whereas out-of-the-money calls and in-the-money puts tend to have low implied
volatilities. (Again, refer to the image at the bottom of the page.)

234
© 2014-2023 AnalystPrep.
Q.1717 Volatility surfaces (the implied volatility as a function of strike price and time to maturity)
are used as pricing tools by traders when dealing with options. Which of the following statements
defines volatility surface CORRECT LY?

A. When volatility smiles and volatility term structures are combined, they produce a
volatility surface.

B. When volatility frowns and volatility term structures are combined, they produce a
volatility surface.

C. When volatility smiles and volatility slopes are combined, they produce a volatility
surface.

D. None of the above.

T he correct answer is A.

Volatility surfaces combine volatility smiles with the volatility term structure to tabulate the
volatilities appropriately.

Q.2860 A foreign currency is valued at $1.73. T he foreign currency has a European call option
market price of 0.135 and a strike price of $1.60. In the US, the risk-free interest rate is 7% per
annum and 16% per annum in the foreign country. Determine the price of a European put option
with a 1-year maturity for the foreign currency.

A. $0.254

B. $0.153

C. $0.548

D. $0.004

T he correct answer is B.

T he price of a European put option with a strike price of $1.60 and 1-year maturity must satisfy the
following equation:
p + S0 e− qT = c + Ke− rT
From the question, we are provided that:
S0 = $1.73, q =16%, T = 1, c = 0.135, K = $1.60 and r = 7%

T herefore:
p + 1.73 × e−0.16×1 = 0.135 + 1.6e−0.07×1
⇒ p = 0.153

235
© 2014-2023 AnalystPrep.
Q.2862 Suppose that a small-cap stock is priced at $0.6560. Suppose further that the price of
European call and put options computed by the Black-Scholes-Merton model are $0.0249 and
$0.0501, respectively. Calculate the market price of a FEB call option if the market price of a FEB
put option is $0.0317.

A. $0.0065

B. $0.0025

C. $0.0337

D. $0.0654

T he correct answer is A.

For the Black-Scholes-Merton model and in the absence of arbitrage opportunities, the put-call parity
satisfies:

P BS + S0 e−qT = CBS + Ke−rt

For the market prices, put-call parity holds when arbitrage opportunities are absent such that:

pMKT + S0e−qT = CMKT + Ke−rt

T he difference between the two equations is:

P BS − P MKT = CBS − c MKT

From the question we have:

CBS = 0.0249, P BS = 0.051 and P MKT = 0.0317

T hus:

0.0501 − 0.0317 = 0.0249 − CMKT

⇒ CMKT = 0.0065

Where,

P BS =put price calculated by the Black-Scholes-Merton model.

CBS =call price calculated by the Black-Scholes-Merton model.

236
© 2014-2023 AnalystPrep.
CMKT =market call price.

P MKT =market put price.

Q.2863 Which of the following condition are necessary for an asset price to have a lognormal
distribution?

I. T he asset should have a varying volatility

II. T he price of the asset should change smoothly with no jumps
III. T he volatility of the asset should be constant

A. I and II

B. I and III

C. II and III

D. All of the above

T he correct answer is C.

Although not usually the case in practical situations, for an asset price to have a lognormal
distribution, the volatility of the asset should not vary and be constant and the change in the asset
price should be smooth without jumps.

237
© 2014-2023 AnalystPrep.
 Reading 78: Fundamental Review of the Trading Book

Q.2362 What’s the number of days of stressed market conditions required for the calculation of the
stressed VaR?

A. 90

B. 150

C. 250

D. 360

T he correct answer is C.

Basel II.5 required banks to calculate a stressed VaR measure in addition to the current measure. As
explained in Chapter 19, this is VaR where calculations are based on the behavior of market variables
during a 250-day period of stressed market conditions. To determine the stressed period, banks were
required to go back through time searching for a 250-day period that would be particularly difficult
for the bank's current portfolio.

Q.2363 In January 2014, the Basel Committee on Banking Supervision issued a consultative document
referred to as the Fundamental Review of the T rading Book (FRT B). T he Basel Committee then
followed requested comments from banks, revised the proposals, and published a final version of the
rules in January 2016. Which of the following is not an acceptable liquidity horizons according to the
BIS January 2016 publication?

A. 10 days

B. 20 days

C. 60 days

D. 140 days

T he correct answer is D.

When implementing Basel I and Basel II.5, banks typically consider one-day changes in market
variables so that a one-day VaR is calculated, and then multiply this VaR by the square root of 10 to
obtain an estimate of the 10-day VaR. FRT B requires the changes to market variables (referred to as
shocks) to be the changes that would take place (in stressed market conditions) over periods of time
that reflect the differing liquidities of market variables. T he periods of time are referred to as
liquidity horizons. Five different liquidity horizons according the BIS January 2016 revised version
are: 10 days, 20 days, 40 days, 60 days, and 120 days.

238
© 2014-2023 AnalystPrep.
Q.2364 On 31st January 2014, a consultative document by the Basel Committee defined an approach
for implementing varying liquidity horizons in which market variables were divided into categories.
What is the criterion used to place the variables into categories?

A. Length of time horizons

B. Severity of change

C. Size of market

D. T ype of product

T he correct answer is A.

One simple approach to implementing varying liquidity horizons would be to use overlapping time
periods. T his type of approach was originally considered by the Basel Committee, but in the January
2014 consultative document, it was rejected in favor of an approach where all calculations are based
on the changes in market variables over 10-day overlapping periods.
Category 1 Variables: Variables with a time horizon of 10 days
Category 2 Variables: Variables with a time horizon of 20 days
Category 3 Variables: Variables with a time horizon of 60 days
Category 4 Variables: Variables with a time horizon of 120 days
Category 5 Variables: Variables with a time horizon of 250 days

Q.2365 Banks sometimes find it difficult to search for past stressed periods using all market variables
because of a shortage of historical data for some of the variables. T he Fundamental Review of
T rading Book (FRT B) allows the stressed period calculations to be based on a subset of market
variables and the results scaled up by the ratio of the expected shortfall for a number of recent
months. What’s the exact number of months as defined by the FRT B?

A. 1

B. 3

C. 6

D. 12

T he correct answer is D.

In practice, banks sometimes find it difficult to search for past stressed periods using all market
variables because of a shortage of historical data for some of the variables. T he FRT B, therefore,
allows the stressed period calculations to be based on a subset of market variables and the results
scaled up by the ratio of the expected shortfall for the most recent 12 months using all market
variables to the expected shortfall for the most recent 12 months using the subset of market
variables. (T he subset of market variables must account for 75% of the expected shortfall.)

239
© 2014-2023 AnalystPrep.
Q.2366 In May 2012, the Basel Committee on Banking Supervision issued a consultative document
referred to as the Fundamental Review of the T rading Book (FRT B). According to FRT B, there are
several categories of market variables. Which of the following is NOT one of them?

A. Equity risk

B. Commodity risk

C. Volatility risk

D. Foreign exchange risk

T he correct answer is C.

Volatility risk does not form part of market variables. T he market variables are divided into a number
of risk categories – (nterest rate risk, equity risk, foreign exchange risk, commodity risk, and credit
risk.

Q.2367 T he Fundamental Review of T rading Book (FRT B) proposes back-testing be done using a VaR
measure calculated over a certain horizon and the most recent 12 months of data. What’s the exact
time horizon used?

A. T hirty-day horizon

B. Seven-day horizon

C. T wo-day horizon

D. One-day horizon

T he correct answer is D.

T he FRT B proposes back-testing be done using a VaR measure calculated over a one-day horizon and
the most recent 12 months of data. (T his is because it is difficult to back-test a 10-day expected
shortfall directly and not possible to back-test stressed VaR or stressed ES.)

240
© 2014-2023 AnalystPrep.
Q.2368 Which of the following presents FRT B's (Fundamental Review of the T rading Book)
proposed change during the measurement of market risk capital?

A. VaR at 99% confidence

B. VaR at 99.99% confidence

C. Expected shortfall with a 97.5% confidence level

D. Expected shortfall with a 99% confidence level

T he correct answer is C.

T he Basel I calculations of market risk capital were based on a value at risk (VaR) calculated for a 10-
day horizon with a 99% confidence level.
T he FRT B is proposing a change to the measure used for determining market risk capital. Instead of
VaR with a 99% confidence level, expected shortfall (ES) with a 97.5% confidence level is proposed.

Q.2369 T he Fundamental Review of T rading Book (FRT B), among other things, defines the type of
instruments which should be put either in the trading or the banking book. Which instruments are
marked-to-market?

A. Instruments in the trading book

B. Instruments in the banking book

C. Instruments in both the banking and the trading books

D. None of the above

T he correct answer is A.

T he FRT B addresses the issue of whether instruments should be put in the trading book or in the
banking book. Roughly speaking, the trading book consists of instruments that the bank intends to
trade. T he banking book consists of instruments that are expected to be held to maturity.
Instruments i n the tradi ng book are mark ed-to-mark et (i .e., reval ued) dai l y while
instruments in the banking book are not.

241
© 2014-2023 AnalystPrep.
Q.2370 According to the Fundamental Review of T rading Book (FRT B), the instruments in the
trading book are subject to:

A. Credit risk capital

B. Market risk capital

C. Operational risk capital

D. Both credit and market risk capital

T he correct answer is B.

Instruments in the banking book are subject to credit risk capital while those in the trading book are
subject to market risk capital.

Q.2371 Basel II.5 introduced the incremental risk charge (IRC), while the Fundamental Review of
T rading Book (FRT B) provides a modification of the IRC by recognizing that for instruments
dependent on the credit risk of a particular company, two types of risk can be identified. T hese are:

A. Downgrade risk and jump-to-default risk

B. Downgrade risk and credit spread risk

C. Credit spread risk and jump-to-default risk

D. Credit spread risk and liquidity risk

T he correct answer is C.

Basel II.5 introduced the incremental risk charge (IRC) to ensure that banks did not reduce capital

requirements by choosing the trading book over the banking book for a credit-dependent instrument.

T he FRT B provides a modification of the IRC. It recognizes that for instruments dependent on the

credit risk of a particular company, two types of risk can be identified:

1. Credit spread risk: T his is the risk that the company's credit spread will change, causing the
mark-to-market value of the instrument to change.
2. Jump-to-default risk: T his is the risk that there will be a default by the company. T ypically
this leads to an immediate loss or gain to the bank.

242
© 2014-2023 AnalystPrep.
Q.2998 Billow Bank has sometimes been finding it difficult to search for past stressed periods using
all market variables. T his fact can be attributed to a shortage of historical data for some of the
variables. How does the Fundamental Review of T rading Book (FRT B) deal with this challenge?

A. T he FRT B can allow the stressed period computations to be based on a subset of market
variables and the results scaled up by the ratio of the expected shortfall for the latest 12
months using all variables to the expected shortfall for the latest 12 months using the subset
of the market variables.

B. An intervention at an early stage by supervisors should prevent capital from falling below
the necessary minimum levels to support a particular bank’s risk characteristics, and a rapid
remedial action should be required in case capital is not maintained or resorted.

C. T he bank should possess a process for assessing the its overall capital adequacy relative
its strategy for capital level maintenance and risk profile.

D. All the above.

T he correct answer is A.

To easily search for past stressed periods using all market variables, the calculation of stressed
periods can be based on a subset of market variables and the results scaled up by the ratio of
expected shortfall for the latest 12 months using all variables to the expected shortfall for the past
12 months using the subset of market variables.

Q.2999 Which of the following reasons best explains why the Fundamental Review of the T rading
Book (FRT B) proposes that back-testing should be done by applying a VaR measure computed over a
one-day horizon and the latest 12 months of data?

A. T his is due to the difficulty in directly backtesting a 10-day expected shortfall and
impossible to back-test stressed VaR or stressed ES.

B. T he events of a contractual default, acceptable remedies, and opportunities for a default

to be cured, should be well defined in the FRT B. T herefore, termination rights should also be
included in agreements.

C. T he reasonability of the proposed limitations compared to the institution’s risks in case of

performance failure by the service provider should be ascertained by the senior
management and the Board of Directors in the institution.

D. None of the above.

T he correct answer is A.

It would be quite difficult for a 10-day expected shortfall to be directly back-tested and also
impossible to back-test stressed VaR or stressed ES.

243
© 2014-2023 AnalystPrep.
Q.4020 In January 2016, the Basel Committee on Banking Supervision (BCBS) issued a revised
framework for Market risk Capital Requirements in part informed by the mass failure of banks
following the 2007-2009 financial crisis. T hat framework has been widely referred to as the
Fundamental Review of the T rading Book (FRT B). Compared to Basel I and Basel II.5, each of the
following is true about the FRT B EXCEPT which is false?

A. Under Basel I and Basel II.5, market risk capital calculations were based on value at risk
(VaR) with a 99.0% confidence level, but the FRT B requires calculations based on expected
shortfall (ES) with a 97.5% confidence level

B. Under Basel I regulations, banks were required to calculate market risk capital based on a
val ue at ri sk calculated for a 10-day hori zon with a 99% confi dence i nterval

C. Under Basel II.5, banks had to add a stressed VaR measure to the current value at risk,
both calculated based on a 10-day horizon.

D. Under Basel I and II.5 market risk capital was based on a 10-day time horizon, but the
FRT B uses five different horizons (10, 20, 40, 60, and 120 days) depending on the liquidity of
the market variable

T he correct answer is C.

Basel II.5 regulations required banks to add a stressed VaR measure to the current value captured
with the 10-day VaR. But unlike the VaR, the stressed VaR used a 250-day period. T he purpose of the
stressed VaR was to measure the behavior of market variables during a 250-day period of stressed
market conditions.

244
© 2014-2023 AnalystPrep.
Q.4021 In the past, a lack of a clear distinction between the regulatory banking book and the trading
book created an opportunity for regulatory arbitrage. T he FRT B attempts to make the distinction
between the trading book and the banking book clearer and less subjective. Which of the following is
FALSE with regard to the FRT B's treatment of the boundary between the two books?

A. T he FRT B establishes a more objective boundary between the regulatory banking and
trading book and severely restricts subsequent movement between the books unless under
extraordinary circumstances

B. Under FRT B, there must be a sincere intent to trade if an asset has to be included in the
trading book

C. FRT B subjects the trading book and banking book to the same set of capital requirements
so as to mitigate regulatory arbitrage

D. In an attempt to further mitigate regulatory arbitrage, FRT B distinguishes two types of

credit risk exposure to a company: credit spread risk and jump-to-default risk

T he correct answer is C.

FRT B recognizes that differences in capital requirements between the trading book and banking
book may give rise to regulatory arbitrage, but it does not propose harmonization of these
requirements as a way to mitigate this risk. Instead, FRT B attempts to establish a clearer boundary
between the two books to make it difficult for banks to misallocate assets.
A, B, and D are all true

Q.4022 Which of the following is not a component of the standardized approach to calculating
regulatory capital for banks under market risk measurement and management.

A. A risk charge calculated using a risk sensitivity approach

B. A residual risk add-on

C. A default risk charge

D. Insolvency risk charge

T he correct answer is D.

Under the standardized approach, the capital requirement is the simple sum of three components:
Risk charges under the sensitivities based method, a default risk charge, and a residual risk add-on.

245
© 2014-2023 AnalystPrep.
Q.4024 Which of the following entities should apply the simplified standardized approach while
calculating their market risk capital?

A. Large systematically important banks (SIBs)

B. Small banks with a low concentration of trading book activity

C. Central banks

D. Banks heavily invested in securitized assets

T he correct answer is B.

In June 2017, the Basel committee put forward a consultative document outlining a simplified version
of the standardized approach for use by small banks. Eligible users are banks with a low
concentration of trading book activity, or those with insufficient infrastructure to successfully
implement the sensitivities-based method.

Q.4025 Under FRT B, the term liquidity horizon represents "the time required to sell a financial
instrument or hedge all its material risks, in a stressed market, without materially affecting market
prices." All of the following liquidity horizons have been proposed by the Basel Committee under the
FRT B framework EXCEPT :

A. 10 days

B. 30-days

C. 60 days

D. 120 days

T he correct answer is B.

Researchers and the Basel Committee on Supervision agree that a uniform 10-day liquidity horizon as
originally proposed in Basel regulations is inappropriate because it is nearly impossible for banks to
exit all risk positions within a 10-day period due to market illiquidity. As such, five different liquidity
horizons have been proposed: 10 days, 20 days, 60 days, 120 days, and 250 days. For
example, the calculation of regulatory capital for a 120-day horizon (essentially 6 months' worth of
trading days) is intended to shield a bank from significant risks while waiting six months to recover
from underlying price volatility

246
© 2014-2023 AnalystPrep.
Q.4026 Under Basel I regulations, banks were required to calculate market risk capital based on a
calculated for a horizon with a confi dence i nterval .

A. Value at risk; 250-day; 99.5%

B. Value at risk; 10-day; 99%

C. Expected shortfall; 10-day; 99.9%

D. Expected shortfall; 10-day; 95%

T he correct answer is B.

Under Basel I regulations, banks were required to calculate market risk capital based on a val ue at
ri sk calculated for a 10-day hori zon with a 99 % confi dence i nterval . T his process generated a
very "current" VaR because the 10-day horizon incorporated a recent period of time, typically one to
four years.

247
© 2014-2023 AnalystPrep.
Q.4027 Under the standardized approach, the capital requirement is the simple sum of three
components: risk charges under the sensitivities based method, a default risk charge, and a residual
risk add-on. Which of the following is not a risk class as defined under the sensitivities-based method.

A. General interest rate risk

B. Foreign exchange risk

C. Commodity risk

D. Funding risk

T he correct answer is D.

Under the first component (risk charges under the sensitivities based method), seven risk classes
(corresponding to trading desks) are defined:

General interest rate risk,

Foreign exchange risk,

Commodity risk,

Equity risk,

Credit spread risk – non securitization,

Credit spread risk – securitization,

Credit spread risk – securitization correlation trading portfolio

For each of these classes, a delta risk charge, vega risk charge, and curvature risk charge are
calculated.

248
© 2014-2023 AnalystPrep.
Q.4028 Under FRT B, the term liquidity horizon refers to:

A. T he time taken to find a willing buyer for a security or commodity

B. T he time taken to successfully execute a sale transaction that guarantees a minimum level
of profit

C. T he time it takes to buy a security or commodity

D. T he time required to sell a security in a stressed market, without materially affecting

market prices

T he correct answer is D.

Under FRT B, the term liquidity horizon represents "the time required to sell a financial instrument
or hedge all its material risks, in a stressed market, without materially affecting market prices."

Q.4029 One of the issues extensively addressed in FRT B has much to do with regulatory
modifications with respect to the trading book and banking book. Which of the following is a major
reason behind these modifications?

A. Internal fraud

B. Default risk

C. Interest rate risk

D. Regulatory arbitrage

T he correct answer is D.

Modifications to the trading book and banking book are largely informed by the need to stump out
regulatory arbitrage. T his is a situation where firms take advantage of the fact that the capital needed
with respect to each book differs. As such, some banks may attempt to allocate instruments to the
book that results in the most favorable capi tal requi rements. For example, some banks are on
record for holding credit-dependent instruments in the trading book because by so doing, they are
subject to less regulatory capital than they would be if they had been placed in the banking book.

249
© 2014-2023 AnalystPrep.
Q.4030 To allocate an asset to the trading book, FRT B proposes two major conditions:

I. bank must be able to trade the asset, and physically manage the associated risks of the
underlying asset on the trading desk.
II. asset must be traded on an exchange
III. T he day-to-day price fluctuations must affect the bank’s equity position and pose a risk to
bank solvency.
IV. T here must be significant default risk on the part of the obligor

A. I and IV

B. II and III

C. I and III

D. IIV only

T he correct answer is C.

FRT B has attempted to make the distinction between the banking book and the trading book clearer
and less subjective. To be allocated to the trading book, the bank must demonstrate more than an
intent to trade. Precisely, some two criteria must be met:

1. T he bank must be able to trade the asset, and physically manage the associated risks of the
underlying asset on the trading desk.
2. T he day-to-day price fluctuations must affect the bank’s equity position and pose a risk to
bank solvency.

250
© 2014-2023 AnalystPrep.
Q.4031 Which of the following approaches should banks use to compute capital with respect to
assets held under a securitization business model?

A. Internal models approach

B. Standardized approach

C. Advanced measurement approach

D. Revised internal models approach

T he correct answer is B.

Basel II.5 introduced the Comprehensive Risk Measure (CRM) charge to cover the risks in
securitized assets such as asset-backed securities and collateralized debt obligations. Under Basel II.5
guidelines, a bank (with regulatory approval) is free to use its own models to determine the CRM
charge.

T hrough the FRT B, the Basel Committee is of the view that this is unsatisfactory because it results

in too much variation in the capital charges calculated by different banks for the same portfolio.

T herefore, FRT B requires banks to use the standardized approach for securitizations.

Q.4032 Richard Glen, FRM, is evaluating market capital requirements for Exim Bank. He starts by
comparing the trading desk’s 1-day static value-at-risk measure (calibrated to the most recent 12
months’ data, equally weighted) at both the 97.5th percentile and the 99th percentile, using two years
of current observations of the desk’s one-day P&L. T he desk experiences 13 exceptions at the 99th
percentile and 32 exceptions at the 97.5th percentile. Based on the results, which of the following
models should Exim bank use to determine its capital needs going forward?

A. Internal models approach

B. Standardized approach

C. Advanced measurement approach

D. Revised internal models approach

T he correct answer is B.

Use of the internal models approach is conditioned on an experience of not more than 12 exceptions
at 99% or not more than 30 exceptions at 97.5% in the most recent 12 month period. Otherwise, all
positions must be capitalized using the standardized approach. Positions must continue to be
capitalized using the standardized method until the desk no longer exceeds the above thresholds over
the prior 12 months.

251
© 2014-2023 AnalystPrep.
Q.4033 A bond portfolio, currently worth $500 million in assets, has a probability of default of 3%.
An analyst evaluates the risk of the portfolio using a 95% value at risk (VaR) and a 95% expected
shortfall (ES). Which of the following is correct?

A. T he VaR shows no loss while the expected shortfall shows a $500 million loss.

B. Both measures will show the same result.

C. T he VaR shows no loss while the expected shortfall shows a $300 million loss.

D. T he VaR shows a loss of $300 million while the expected shortfall shows no loss.

T he correct answer is C.

T he VaR measure would show a $0 loss because the probability of default is less than 5%. A 3%
probability of default implies that three out of five times, in the tail, the portfolio will experience a
total loss. T he potential loss is $300 million (= 3/5 × $500million) .

Q.4034 During a workshop set up by a banking regulator regarding market capital calculations, an
intern makes the following statements regarding the differences between Basel I, Basel II. 5, and the
Fundamental Review of the T rading Book (FRT B). Which statement is i ncorrect ?

A. Both Basel I and Basel II. 5 require calculation of VaR with a 99% confidence interval.

B. FRT B requires the calculation of expected shortfall with a 97.5% confidence interval.

C. FRT B requires adding a stressed VaR measure to complement the expected shortfall
calculation.

D. T he 10-day time horizon for market risk capital proposed under Basel I incorporates a
recent period of time, which typically ranges from one to four years.

T he correct answer is C.

Under FRT B proposals banks no longer have to combine the 10-day VaR and the 250-day stressed
VaR risk measures. Instead, they are required to calculate capital based exclusively on expected
shortfall using a 250-day stressed period. However, banks have retained the freedom to self-select a
250-day window of exceptionally difficult financial stress.

252
© 2014-2023 AnalystPrep.
Q.4035 Which of the following risks is specifically recognized by the incremental risk charge(IRC)?

A. Jump to default risk

B. Interest rate risk

C. Foreign exchange risk

D. Equity price risk

T he correct answer is A.

T he two types of risk recognized by the incremental risk charge are: (1) credit spread risk, and (2)
jump-to-default risk. Credit spread risk can be addressed by calculating the expected shortfall, while
Jump-to-default risk is measured by 99% VaR and not 97.5% expected shortfall.

253
© 2014-2023 AnalystPrep.