FRM Test 12 - Topic - Book 2 - 4

FRM Part – 1 – Test ID – 0012 – Topic - Book 2 & 4 – Questions - 100
Answer Sheet
Probabilities
1. Which of the following can be categorized as continuous random variables?
I. Stock indices
II. The weight of 20 FRM candidates
III. Biannual share dividends received over a 10-year period
IV. The income of citizens in a given district
V. The annual number of FRM exam candidates in the last 10 years
A. I, III, and V
B. I, II, and III
C. II and V
D. All the above
The correct answer is: B).
A continuous random variable is a variable that has infinite possible outcomes, even though
lower and upper bounds exist. It differs from a discrete random variable which takes on only a
countable number of values. As an example, between 2433 and 2434, the S&P 500 index can
take on 2433.00, 3433.23, 2433.89, 2433.07, etc. On the other hand, the number of FRM
candidates in any given year can only take on an integer/whole number value.
2. Consider the following probability function for a discrete random variable X:
X = {10, 20, 30, Y,}, P(x) = x/100, otherwise P(x) = 0
Find the value of Y.

A. 4
B. 40
C. 50
D. 5
There are two key properties that any probability function must meet. These are:
I. 0 ≤ P(x) ≤ 1
II. ∑P(x) = 1
Therefore,
10/100 + 20/100 + 30/100 + Y/100 = 1
60/100 + Y/100 = 1
60 + Y = 100
Y = 40
3. A zero-coupon bond with a notional value of $100 and price x has the following probability
density function:
f(x) = x/5000 for 0 ≤ x ≤ 100
Determine the probability that the price of the bond is between $80 and $90.
A. 0.0017
B. 0.81
C. 0.17
D. 0.64
The correct answer is: C).
To find the required probability, simply integrate the given probability density function between
80 and 90.
=80∫ 90x/5,000dx
= [x2/10,000]x = 90 - [x2/10,000] x = 80
= 8100/10,000 – 6400/10,000
= 0.81 – 0.64
= 0.17 or 17%
4. Given the following probability density function for a discrete random variable X, determine
F(30).
X = {10, 20, 30, 40}, P(x) = x/100

A. 0.2
B. 0.3
C. 0.7
D. 0.6
The correct answer is: D).
The cumulative distribution function (cdf) F(x) defines the probability that a random variable X,
and assumes a value equal to or less than a specified value, x. As such:
F(30) = P(X = 30)
= P(X = 10) + P(X = 20) + P(X = 30)
= 10/100 + 20/100 + 30/100
= 0.6 or 60%
5. The probability of an increase in the annual dividend paid out to shareholders of ABC Limited
is 0.4. The probability of an increase in share price given an increase in dividends is 0.7.
Determine the joint probability of an increase in dividends and an increase in share price.
A. 0.28
B. 0.14
C. 0.72
D. 0.3
The correct answer is: A).
Let:
A be the event that the dividend is increased and,
B be the event that the share price increases
Therefore, P(A) = 0.4 and P(B | A) = 0.7
The joint probability of an increase in dividends and an increase in share price is P(B ᴖ A)
The multiplication rule of probability states that:

P(B | A) = P(B ᴖ A)/P(A)
Hence P(B ᴖ A) = P(B | A) * P(A)
= 0.7 * 0.4
= 0.28 or 28% (Note that P(A ᴖ B) = P(B ᴖ A))
6. Two events are said to be independent and mutually exclusive if:

A. The events cannot occur at the same time
B. The occurrence of one of the events does not affect the probability of the occurrence of the
other event
C. The occurrence of one of the events does not affect the probability of the occurrence of the
other event and the events cannot occur at the same time
D. The occurrence of one of the events means the second event is certain to occur
Independent events are those whose occurrences are uncorrelated. The occurrence of one
event does not in any way affect the chances of the other event occurring. In addition, if two
events are mutually exclusive, it means they cannot occur at the same time. For example, if a
fair coin is tossed, it’s impossible to obtain a head and a tail at the same time since the two
possible outcomes are mutually exclusive.
7. A financial risk manager exam candidate is asked two questions. The probability that she
gets the first question correct is 0.4 and the probability that she gets the second question correct
is 0.5. Given that the probability that she gets both questions correct is 0.2, determine the
probability that she gets either the first or the second question correct.
A. 0.9
B. 0.7
C. 0.1
D. 0.4
Let A = {gets first question right} and B = {gets second question right}
Therefore, P(A) = 0.4, P(B) = 0.5, and P(A ᴖ B) = 0.2
We want to determine P(A ᴗ B);
The addition rule states that P(A ᴗ B) = P(A) + P(B) – P(A ᴖ B)
Hence, P(A ᴗ B) = 0.4 + 0.5 – 0.2 = 0.7
8. An empirical study of ABC stock listed on the New York Exchange reveals that the stock has
closed higher on one-third of all days in the past few months. Given that up and down days are
independent, determine the probability of ABC stock closing higher for six consecutive days.
A. 0.17
B. 0.0137
C. 0.03704
D. 0.00137
From the information above, we can establish that the probability of closing higher = 1/3
Using independence, the probability of 6 consecutive “highs” = 1/3 * 1/3 * 1/3 * 1/3 * 1/3 * 1/3
= 1/729
(The calculation above follows from the fact that if A and B are independent events, then P(A ᴖ
B) = P(A) * P(B).)
9. A fruit juice shop allows customers to choose apple juice, mango juice or passion juice. The
probability of a customer ordering passion juice is 0.45, mango juice and apple juice 0.19,
passion juice and mango juice 0.15, passion juice and apple juice 0.25, passion juice or mango
juice 0.6, passion juice or apple juice 0.84, and 0.9 for at least one of them.
Find the probability that a customer orders all the three juices.
A. 0.64
B. 0.1
C. 0.3
D. 0.25
Let:
A be the event that a customer chooses/orders apple juice
M be the event that a customer chooses mango juice
S be the event that a customer chooses passion fruit
We can easily establish that:

P(S) = 0.45, P(M ᴖ A) = 0.19, P(M ᴖ S) = 0.15, P(A ᴖ S) = 0.25, P(M ᴗ S) = 0.6, P(A ᴗ S) = 0.84,
P(A ᴗ M ᴗ S) = 0.9
We need to determine P(A ᴖ M ᴖ S):

Borrowing from the addition rule with three sets,
P(A ᴗ M ᴗ S) = P(A) + P(M) + P(S) - P(M ᴖ A) - P(M ᴖ S) - P(A ᴖ S) + P(A ᴖ M ᴖ
S)…………………..equation (I)
P(M ᴗ S) = P(M) + P(S) – P(M ᴖ S),
P(M) = 0.6 + 0.15 – 0.45 = 0.3
Similarly, P(A ᴗ S) = P(A) + P(S) - P(A ᴖ S)

P(A) = 0.84 – 0.45 + 0.25 = 0.64
Therefore applying equation (I),

0.9 = 0.64 + 0.3 + 0.45 – 0.19 – 0.15 – 0.25 + P(A ᴖ M ᴖ S)
Which gives us P(A ᴖ M ᴖ S) = 0.1
10. A random variable X with an exponential distribution has the following probability density
function:
FX(x) = λe- xy
You are required to determine the probability density function of the random variable X, given
that Y = X2
A. 1 – e-λy^1/2
B. – e-λy^1/2
C. 1/2 λy- ½ e- λy^1/2
D. 1
From first principles, it can be shown that:

FY(y) = P(Y ≤ y) = P(X2 ≤ y) = P(X ≤ y1/2)
Applying integration:
P(X ≤ y1/2) = 0∫y^1/2(λe-xy)dx = [-e-λx]x = y^1/2 - [-e- λx]x = 0 = 1 – e-λy^1/2
FY(y) = F’Y(y) = 1/2 λy- ½ e-λy^1/2
11. During a lottery, 400 names are fed into a computer program. Five of the names are
identical. If a name is drawn from the program at random, what is the probability that one of
these 5 names will be drawn?
A.0.0125
B. 0.25
C. 0.0025
D. 0.0625
P(name 1 ᴗ name 2 ᴗ name 3 ᴗ name 4 ᴗ name 5) = 1/400 + 1/400 + 1/400 + 1/400 + 1/400 =
0.0125
12. If two events are not independent, the joint probability of A and B, P(A ᴖ B) is equal to:
A. P(A | B)/P(B)
B. P(A | B) * P(B)
C. P(A) * P(B)
D. None of the above
If two events are not independent, then the occurrence of one of the events affects the chances
of occurrence of the other event.
As such, P(A | B) = P(A ᴖ B)/P(B)

Which can be rewritten to make P(A ᴖ B) the subject of the formula.
If A and B were independent, then P(A ᴖ B) = P(A) * P(B)
13. Use the following incomplete probability matrix to compute the joint probability of a poor
economy and an increase in interest rate:
Interest rates
Increase No increase
Economy Good 20% 10%
Normal 30% 20%
Poor X 10%
A. 0.2
B. 0.3
C. 0.1
D. 0.05
All the joint probabilities must add up to 1. Therefore,

X = 100% - 20% - 30% - 10% - 20% - 10%
= 10%
14. Calculate the probability that a stock return is either below -5% or above 5%, given:
P(R < -5%) = 16%

P(R > +5%) = 18%
A. 0.02
B. 0.32
C. 0.34
D. 0.16
Since the two events are mutually exclusive, the return cannot be below -5% or above 10% at
the same time. Therefore, the answer is simply 16% + 18% = 34%
Basic Statistics
15. Compute the sample standard deviation given the following sample data:
∑x = 31,353
n = 100
∑x2 = 10,687,041
A. 8654.64
B. 93
C. 313.53
D. 315
The formula for calculation of sample variance is s2 = 1/(n – 1)[∑x2 – n x̄2]
From the data, x̄ = 31353/100 = 313.53
s2 = 1/99[10,687,041 – 100 * 313.532] = 8654.64
The sample standard deviation = √s2 = √8654.64 = 93
16. On Tuesday, an insurance company receives a total of 10 claims for automobile policies.
After first round assessment, it’s found that the mean claim amount is $426 while the standard
deviation is 112. On Tuesday, the chief claims analyst authorizes the removal of one of the
claims for $545 from the list on grounds that it’s fraught with fraud. Compute the standard
deviation for the remaining set of 9 claims.
A. 110.2
B. 12145.2
C. 421.8
D. 420
∑x = the total of the original set of 10 claims = 426 * 10 = 4260
After removing the claim worth $545, the new value of ∑x = 4260 – 545 = 3,715
Thus, the new mean = 3715/9 = $412.8
Since s2 = 1/(n – 1)[∑x2 – nx̄2], using data for all the ten claims,
1122 = 1/9[∑x2 – 10 * 4262]
Making ∑x2 the subject of the formula,

∑x2 = 1122 * 9 + 10 * 4262 = 1,927,656
Removing the fraudulent claim gives,

∑x2 = 1,927,656 – 5452 = 1,630,631
Now, using the data for the remaining 8 claims,

S2 = 1/(9 – 1)[1,630,631 – 9 * 412.82]
= 12,145.2
Therefore, s = √12,145.2 = 110.2
17. The following table presents the probability distribution of the earnings per share (EPS) for a
certain company:
Probability EPS Interest rates Beta
20% $1.5 21 1.15
10% $2.0 20% 10%
30% $1.3 20% 1.25
40% $1.2 10%
Compute the expected earnings per share.

A. 2
B. 1.5
C. 1.37
D. 1.2
EPS = E(X) = ∑P(xi)xi

= 0.2 * 1.5 + 0.1 * 2.0 + 0.3 * 1.3 + 0.4 * 1.2
= 1.37
18. A renowned economist has calculated that the Canadian economy will be in one of 3
possible states in the coming year: Boom, Normal, or Slow. The following table gives the returns
of stocks A and B under each economic state.
State Probability(state) Return for stock A Return for stock B

Boom 40% 12% 18%
Normal 35% 10% 15%
Slow 25% 8% 12%
Which of the following is closest to the covariance of the returns for stocks A and B?
A. 0.103
B. 0.0001734
C. 0.1545
D. 0.0003765
Cov(A, B) = ∑P(s) * [RA – E(RA)] * [RB – E(RB)]
First, you have to determine the expected return for every stock:
E(RA) = 0.4 * 0.12 + 0.35 * 0.1 + 0.25 * 0.08 = 0.103
E(RB) = 0.4 * 0.18 + 0.35 * 0.15 + 0.25 * 0.12 = 0.1545
State P(S) RA RB P(S) * [RA – E(RA)] * [RB

– E(RB)]
Boom 0.4 0.12 0.18 0.4 * [0.12 – 0.103] * [0.18 – 0.1545] = 0.0001734
Normal 0.35 0.1 0.15 0.35 * [0.1 – 0.103] * [0.15 – 0.1545] = 0.000004725
Slow 0.25 0.08 0.12 0.25 * [0.08 – 0.103] * [0.12 – 0.1545] = 0.0001984
Cov(A, B) = 0.0001734 + 0.000004725+ 0.0001984
= 0.0003765
19. Which of the following statements is NOT true regarding the correlation coefficient?
A. The correlation coefficient measures the strength of the linear relationship between two
random variables
B. The correlation coefficient has no units
C. The correlation coefficient ranges from 0 to +1
D. Random variables with a correlation of +1 are said to be perfectly correlated
In finance, the correlation coefficient attempts to measure the degree to which two random
variables, say, returns for different stocks move in relation to each other. The correlation
coefficient always lies between -1 and +1. A positive value indicates that the random variables
move in the same direction i.e. if an increase (decrease) is recorded in one variable, we expect
an increase (decrease) in the other variable, which can either be proportionate or
disproportionate depending on the value of the correlation.
20. A discrete random variable Y has probability function given by:

Y 0 1 2
P(Y = y) 0.3 0.6 0.1
Calculate Var(Y).
A. 0.2
B. 0.36
C. 0.8
D. 1
The variance of any given random variable is given by:

Var(Y) = E(Y2) - E2(Y)
E(Y) = ∑YP(Y =y) = 0 * 0.3 + 1 * 0.6 + 2 * 0.1 = 0.8
E(Y2) = ∑Y2P(Y =y) = 02 * 0.3 + 12 * 0.6 + 22 * 0.1 = 1
Therefore, Var(Y) = 1 - 0.82 = 0.36
21. The mean height of female FRM exam candidates over the years is 1.671m while that of
male candidates is 1.757. Given that the mean height of ALL of the exam candidates is 1.712m,
calculate the percentage of the candidates who are female:
A. 0.523
B. 0.46
C. 0.087
D. 0.477
Let F be the proportion of females.
Applying the idea of a weighted mean,

1.671F + 1.757(1 – F) = 1.712
1.671F + 1.757 – 1.757F = 1.712
1.757 -1.712 = 1.757F - 1.671F
0.045 = 0.086F
F = 0.523 or 52.3%
22. Two random variables X and Y are such that V[X] = 4V[Y] and Cov[X,Y] = V[Y]
Let E = X + Y and F = X – Y
Find Cov[E, F]
A. V[Y] – V[X]
B. Cov[X,Y]
C. V[Y]
D. 3V[Y]
Cov[E,F] = Cov[X + Y,X – Y]

= Cov[X,X] – Cov[X,Y] + Cov[Y,X] – Cov[Y,Y]
= V[X] – V[Y]
= 4V[Y] – V[Y]
= 3V[Y]
Logic applied:
I. Given a random variable X, the covariance between X and itself is simply its variance
II. Cov[X,Y] = Cov[Y,X]
23. Let Y be a discrete random variable with the following probability distribution:
Y 0 1 2 3
P(Y = y) 0.3 0.2 0.4 0.1
Determine the variance of X, where X = 2Y + 10

A. 1.7
B. 2.7
C. 4
D. 14
V[X] = V[2Y + 10]

= 4V[Y]
V[Y] = E[Y2] – E2Y
= 0.3 * 02 + 0.2 * 12 + 0.4 * 22 + 0.1 * 32 – [0.3 * 0 + 0.2 * 1 + 0.4 * 2 + 0.1 * 3] 2

= 2.7 – 1.7
=1
Hence V[X] = 4V[Y] = 4 * 1 = 4
Logic applied:
I. V[aX] = a2V[X]
II. V[X ± a] = V[X]
Where X is a random variable and a is a constant.
24. Which of the following best describes the concept of skewness in statistics?
A. The degree to which a distribution is symmetric about its mean
B. The degree to which a distribution is nonsymmetric about its median
C. The degree to which a distribution is nonsymmetric about its mean
D The degree to which a random variable spreads around its mean
Skewness in statistics describes the asymmetry from the normal distribution in a set of data.
Such a dataset differs from a normal curve which is bell-shaped and perfectly symmetrical. In
layman’s language, a symmetrical curve can be divided into two equal halves with the mean at
the middle. When this is not possible, the curve (and the underlying data) is said to be skewed.
A distribution can either be positively or negatively skewed, depending on where there’s a
higher concentration of data points.
25. Which of the following is incorrect about kurtosis?

A. Excess kurtosis is a measure relative to the uniform distribution, which has a kurtosis of 3
B. Excess kurtosis that’s negative indicates a platykurtic distribution
C. Excess kurtosis that’s positive indicates a leptokurtic distribution
D. The normal distribution has a kurtosis equal to 3
Kurtosis basically measures the peakedness of a distribution. Data sets with a high kurtosis
tend to have many data points at the tails (outliers). Kurtosis is measured relative to the normal
distribution, which has a kurtosis of exactly 3.
26. Mary, FRM, is tasked with analyzing the returns of two different assets – A and B. She finds
that the two assets have the same mean, variance, and skewness, but A has a higher kurtosis
than B. Which of the following statements is most likely true?
A. Asset A is riskier than asset B
B. Asset B is riskier than asset A
C. Both assets are highly profitable
D. Assets A and B will earn equal returns in the long term
In finance, Kurtosis affects the riskiness of an asset. The asset with a higher kurtosis is
considered riskier than another one with a lower kurtosis. The underlying logic is that a high
kurtosis indicates a high number of outliers, meaning that the return for such an asset is highly
variable, and therefore highly risky.
27. The following are measures of variability, EXCEPT:

A. Variance
B. Standard deviation
C. Range
D. Median
The mean, median, and the mode are all measures of central tendency – all of them attempt to
describe data by identifying a central position. Measures of variation such as variance describe
the spread of the data around the mean.
28. Two stocks, X and Y, have a correlation of 0.50. Stock Y’s return has a standard deviation of
0.26. Given that the covariance between X and Y is 0.005, determine the variance of returns for
stock X
A. 0.13
B. 0.00148
C. 0.0385
D. 0.0148
Correlation between X and Y,

Corr(X,Y) = Cov(X,Y)/(sX * sY)
0.50 = 0.005/(σX * 0.26)
0.13σX = 0.005
σX = 0.0385
V(X) = 0.03852 = 0.00148
29. Which of the following best describes the concept of an unbiased estimator?
A. One for which the accuracy of the parameter estimate increases as the sample size
increases
B. One that has the least variance compared to all other estimators
C. One for which the accuracy of the parameter estimate increases as the sample size
decreases
D. One for which the expected value of the estimator is equal to the value of the parameter
being estimated
If x̄ is an unbiased estimator of μ, then the expected value of x̄ is equal to μ.
Distributions
30. During a disease outbreak, the probability of surviving after infection is 60%. Determine the
probability that at least 8 out of a group of 9 infected persons will survive?
A. 0.7
B. 0.07
C. 0.007
D. 0.06
We note the following:

I. That the infection of any single individual is independent of all other infections for other
individuals.
II. The trials are identical (Probability of survival is 60% every time)
Therefore, the number of survivors takes on a binomial distribution with n = 9 and θ = 0.6
If X stands for the number of survivors, then:

P(X ≥ 8) = P(X=8 or 9) = 9C8 * 0.68 * 0.4 + 9C9 * 0.69
= 9 * 0.017 * 0.4 + 1 * 0.01
= 0.070 or 7%
31. Luke Friday, FRM, runs a consultancy firm that offers investment advice to clients around
Los Angeles. The number of clients the firm receives in a month is distributed as a Poisson
variable with a mean of 2. What is the probability that the firm receives exactly 30 new clients in
a year, assuming every client is independent?
A. 0.025
B. 0.0363
C. 0.24
D. 0.00363
We note the following:

I. New clients are received randomly at a rate of 2 per unit time (month)
II. Each event is independent
These observations confirm the Poisson distribution.
Since the question asks us to find a yearly probability, we must convert the monthly Poisson
rate to an equivalent yearly rate.
Poi(2) monthly distribution is equivalent to Poi(24)
Now, if X stands for the number of clients received,

P(X = x) = [exp(-λ) λ x]/x!
P(X = 30) = [exp(-24) 2430]/30!
= 0.0363
32. Which of the following is NOT true regarding the normal distribution?
A. It’s completely described by its mean, μ, and variance, σ2
B. Its skewness = 3 and kurtosis = 0
C. A linear combination of two normally distributed variables also has a normal distribution
D. The probabilities of extreme events (those further above and below the mean) continually get
smaller but extend infinitely without going to zero
Statement B is false but its converse is true: The normal distribution has skewness = 0 and
kurtosis = 3. In fact, the kurtosis of other distributions is measured relative to 3, which is the
kurtosis of the normal distribution.
33. Consider the following events:
I. Throwing a fair, six-sided die

II. The rate at which customers walk into a banking hall per day
III. The score of 50 FRM exam candidates in a mock test
IV. Tossing a coin
V. Picking of an orange from a basket containing 10 equally sized oranges
Which of the events above exhibit uniform distributions?

A. I, IV, and V
B. I and II only
C. II and III only
Under the uniform distribution, ALL outcomes are equally likely i.e., they have equal
probabilities of occurrence. For example, if we were to throw a fair die, each of numbers 1 to 6
would have a probability of 1/6. Similarly, a head (or a tail) occurs with probabilities of 0.5 when
a coin is tossed.
34. The probability that a patient suffering from typhoid will be treated successfully is 0.8. 40
patients are subjected to treatment. Determine the expected value of the number of patients
who are treated successfully.
A. 7
B. 28
C. 8
D. 32
This question tests the knowledge of the mean of the binomial distribution (n, θ)
The expected number of cured patients = E(X) = nθ = 40 * 0.8 = 32
Note that V(X) = nθ (1 – θ)
35. The rate of registration for the FRM exam by candidates takes on a Poisson distribution with
mean λ. Which of the following statements is correct?
A. Mean equals the standard deviation
B. Mean equals the variance
C. Median equals the variance
D. Median, mean and variance are all equal
An interesting fact about the poisson distribution is that the mean equals the variance.
36. A vehicle repairs assembly has a total of 100 jerks and other repair work machines in
constant use. The probability of a machine breaking down during a given day is 0.004. There
are days when none of the machines break down. However, during some days, one, two, three,
four, or more machines break down. Calculate the probability that fewer than 3 machines break
down during a particular day.
A. 0.007726
B. 0.6698
C. 0.9923
D. 0.269
The number of breakdowns takes on a binomial distribution with n = 100 and θ = 0.004
“Fewer than 3” implies 0, or 1, or 2 machines break down
Therefore,
P(fewer than 3) = P(0 breakdowns) + P(1 breakdowns) + P(2 breakdowns)
= 100C0 * 0.0040 * 0.996100 + 100C1 * 0.0041 * 0.99699 + 100C2 * 0.0042 * 0.99698
= 0.6698 + 0.2690 + 0.05347
= 0.9923
37. In the standard normal distribution, what do z-scores represent?

A. Scores below and above the mean in units of the standard deviation of the distribution
B. Scores below and above the variance in units of the standard deviation of the distribution
C. Scores above the mean in units of the standard deviation of the distribution from the mean
D. Scores below and above the mean in units of the standard deviation of the distribution from
the mean
The z-score is calculated as (X – μ)/ σ
It therefore gives the location of raw scores above and below the mean in units of the standard
deviation of the distribution from the mean.
38. The random variable X denotes (in units of $100,000) the size of loss per project incurred in
a particular investment company. In addition, assume that X follows a chi-square distribution
with 2 degrees of freedom. A risk manager randomly chooses two such projects and further
assumes that their corresponding losses are independent of each other. Calculate the mean
and variance of the total loss from the two projects.
A. Mean = 400,000; variance = 8 * 1010
B. Mean = $400,000; variance = $8 * 1010
C. Mean = 400,000; variance = 800,000
D. Mean = $400,000; variance = $800,000
It’s imperative to note the following: A random variable that follows the chi-square distribution
has a mean of n and a variance of 2n, where n is the number of degrees of freedom. We could
approach the question from two different perspectives:
1. Summation of independent variables

Since n = 2, E(Xi) = 2, and Var(Xi) = 4
X1 and X2 are independent, which means:
E(X1 + X2) = 4, and Var(X1 + X2) = 8
Therefore, for the total loss from the two projects,
Mean = 4 * 100,000 = $400,000 while Variance = $8 * 1010 (we have to square up the dollar -
$2)
2. Summation of two chi-square random variables

If A and B are two chi-square random variables with m and n degrees of freedom respectively,
then the sum of A and B is ALSO a chi-square variable with m + n degrees of freedom.
Therefore, X1 + X2 = X4
Working out the problem with this result (X4) gives the same values as above.
39. Which of the following best describes the central limit theorem?
A. When the sample size is large, the sum of independent and identically distributed (i.i.d.)
random variables are normally distributed
B. The sum of n independent and identically distributed random variables approaches the
normal distribution as n becomes large
C. For simple random samples of size n from a population with mean μ and finite variance σ2,
the sampling distribution approaches the normal distribution with mean μ and variance σ2/n, as
the sample size becomes large
D. For simple random samples of size n from a population with mean µ and finite variance σ2,
the sampling distribution of the sample mean approaches the normal distribution with mean μ
and variance σ2/n, as the sample size becomes large
The Central Limit Theorem states that the sampling distribution of the sample means
approaches a normal distribution as the sample size gets larger — no matter what the shape of
the population distribution. This fact holds especially true for sample sizes over 30. All this is
saying is that as you take more samples, especially large ones of size 30 or more, your graph of
the sample means will look more like a normal distribution. Please note that it's the sample
mean that's normally distributed, not the sample itself. This is why option C is incorrect. This
handy result considerably simplifies computation of probabilities and construction of statistical
hypothesis. So long as we have sample statistics, we can draw relevant conclusions about the
actual population regardless of the population’s distribution, provided n is sufficiently large (n is
usually taken to be ≥ 30).
40. The marketing department of a large mutual fund estimates that 82% of all new employees
put on probation for the first year eventually get fully employed. During a recent recruitment
drive, a total of 280 new employees were recruited. Approximate the probability that at least 240
of these will eventually earn themselves permanent roles in the company after one year.
A. 0.062
B. 0.18
C. 0.9382
D. 0.82
If we let X to be the number of employees who get fully employed:

X follows a binomial distribution with mean (θ) 0.82 and n = 280 i.e. X∿bin(280, 0.82)
We want to find P(X ≥ 240)
It would take a lot of time to compute P(X = 240, X = 241 … X = 280). Thanks to the central limit
theorem, we can approximate the binomial distribution in terms of the normal distribution,
whereby:
X∿bin(280, 0.82) ≈ N(229.6, 6.432)
where mean becomes nθ while variance becomes nθ(1 – θ)
Now, P(X ≥ 240) = P(X > 239.5) applying continuity correction as a result approximating a
discrete distribution (binomial) with a continuous distribution (normal)
= P[Z > (239.5 – 229.6)/6.43] = P(Z > 1.54) = 1 – 0.93822 = 0.062 (read off the normal
distribution table)
41. The normal distribution and the lognormal distribution are related in such way that:
A. If a random variable X follows a lognormal distribution, ln X is normally distributed
B. If a random variable X follows a normal distribution, ln X is said to have a lognormal
distribution
C. The mean and variance of a lognormal distribution are twice that of the normal distribution,
provided the value of n is the same
D. The mean and variance of the normal distribution are twice that of the lognormal distribution,
provided the value of n is the same
A random variable X follows a lognormal distribution if its natural logarithm, ln X, is normally

distributed. In layman’s language (for easy understanding), you can view the term “lognormal”
as “the log is normal”.
42. A motor vehicle production company based in California is assembling its first batch of fully
electric cars. After inspecting about 100 newly assembled units, engineers establish that 40 of
them have mechanical defects. While some units have no defects, others have one, two, or
more defects. Assume that the distribution of mechanical defects follows a Poisson distribution.
Drawing on the first 100 units produced, how many cars, out of every 10,000 units assembled,
would we expect to have at least one defect?
A. About 330
B. About 0.330
C. About 3,300
D. About 1250
Let’s use X to denote the number of defects in a car.

X∿Poi(40/100) i.e. λ = 0.4
P(X is at least 1) = 1 – P(X = 0)

= 1 – exp(-0.4) = 0.330
This is the probability of at least 1 defect in a car. Therefore, for every 10,000 cars, we would
expect 0.330 * 10,000 = 3,300 units to have one or more defects.
43. The F-distribution and the Chi-square distribution have glaring similarities. Which of the
following is not accurate?
A. Both are asymmetrical
B. Both have a bound equal to zero on the left
C. Their means are always less than their standard deviations
D. They are defined by the number of degrees of freedom
There exists no consistent relationship between mean and standard deviation in either the F- or
the chi-square distribution.
44. Insurance claims in a certain class of business are modeled using a normal distribution with
mean $3,000 and standard deviation $400. Calculate the probability that the next claim received
will exceed $3,500.
A. 0.8944
B. 0.25
C. 0.75
D. 0.1056
X∿N(3000, 4002)
P(X > 3500) = P[Z > (3500 – 3000)/400]

= P(Z > 1.25)
= 1 –P(Z < 1.25)
=1 - 0.8944
= 0.1056
Bayesian Analysis
45. The punctuality of filing tax returns has been investigated by considering a number of
citizens in different geographical regions. In the sample, 60 % of respondents were from Africa,
20% Europe, and 20% South America. The probabilities of late filing of returns in Africa, Europe,
and South America are 45%, 15%, and 20% respectively.
If a late submitter is picked at random from the area under study, what is the probability that
they are from Africa?
A. 0.7941
B. 0.0794
C. 0.34
D. 0.27
Let ‘A’ be the event that an individual chosen at random comes from Africa. Let ‘E’ and ‘S’ have
similar definitions for Europe and South America respectively.
Define ‘L’ as the event that an individual chosen at random submits tax returns late.
Now, we wish to determine P(Africa | Late) = P(A | L)
Applying Bayes Theorem,

P(A| L) = (P(A) * P(L | A))/ [P(A) * P(L | A) + P(E) * P(L | E) + P(S) * P(L | S)]
= 0.6 * 0.45/[(0.6 * 0.45) + (0.2 * 0.15) + (0.2 * 0.20]
= 0.27/[0.27 + 0.03 + 0.04]
= 0.7941
46. A financial risk manager has three routes to get to the office. The probability that she gets to
the office on time using routes X, Y, and Z are 60%, 65%, and 70%. She does not have a
preferred route and is therefore equally likely to choose any of the three routes. Calculate the
probability that she chose route Z given that she arrives to work on time.
A. 0.359
B. 0.233
C. 0.216
D. 0.2
Define X to be the event “chooses route X” Let Y and Z have similar definitions.
Define O to be the event that she arrives on time
We wish to determine P(Z | O). Then:

P(Z | O) = P(Z) * P(O | Z)/ [P(Z) * P(O | Z) + P(Y) * P(O | Y) + P(X) * P(O| X)]
= (1/3 * 0.7)/[(1/3 * 0.7) + (1/3 * 0.65) + (1/3 * 0.6)
= 0.2333/(0.2333 + 0.2167 + 0.2)
= 0.3589
47. Common Text for Questions 357, 358, 359, and 360
A life assurance company insures individuals of all ages. A manager compiled the following
statistics of the company’s insured persons:
Age of Mortality (Probability of Portion of company’s

insured death)[arbitrary] insured persons
16-20 0.04 0.1

21-30 0.05 0.29
31-65 0.10 0.49
66-99 0.14 0.12
If a randomly selected individual insured by the company dies, calculate the probability that the
dead client was age 16-20.
A. 0.04
B. 0.048
C. 0.046
D. 0.047
Define the following events:

B = Event of death
B1= Event the insured’s age is in the range 16-20
We wish to determine P(B1 | B)

P(B1 | B) = (P(B1) * P(B | B1))/[ P(B1) * P(B | B1) + (P(B2) * P(B | B2) + (P(B3) * P(B | B3) + (P(B4)
* P(B | B4)]
= (0.1 * 0.04)/[(0.1 * 0.04) + (0.29 * 0.05) + (0.49 * 0.1) + (0.12 * 0.14)
= 0.004/(0.004 + 0.0145 + 0.049 + 0.0168)
= 0.04745 or 4.7%
48. A life assurance company insures individuals of all ages. A manager compiled the following

16-20 0.04 0.1
21-30 0.05 0.29

31-65 0.10 0.49
66-99 0.14 0.12
If a randomly selected individual insured by the company dies, calculate the probability that the
dead client was in age range 21-30.
A. 0.172
B. 0.04
C. 0.168
D. 0.145
P(B2 | B) = (P(B2) * P(B | B2))/[ P(B2) * P(B | B2) + (P(B1) * P(B | B1) + (P(B3) * P(B | B3) + (P(B4) *
P(B | B4)]
= (0.29 * 0.05)/[(0.29 * 0.05) + (0.1 * 0.04) + (0.49 * 0.1) + (0.12 * 0.14)
= 0.0145/(0.0145 + 0.004 + 0.049 + 0.0168)
= 17.2%
49.
A life assurance company insures individuals of all ages. A manager compiled the following

16-20 0.04 0.1
21-30 0.05 0.29
31-65 0.10 0.49
66-99 0.14 0.12
Compute the probability that the dead client was in age range 31-65.
A. 0.58
B. 0.172
C. 0.168
D. 0.047
* P(B | B4)]
= (0.49 * 0.10)/[0.49 * 0.10 + 0.1 * 0.04 + 0.29 * 0.05 + 0.12 *0.14]
= 0.049/(0.049 + 0.004 + 0.0145 + 0.0168)
= 58%
50. A life assurance company insures individuals of all ages. A manager compiled the following

16-20 0.04 0.1
21-30 0.05 0.29
31-65 0.10 0.49
66-99 0.14 0.12
Calculate the probability that the dead client was between 66 and 99 years.
A. 0.047
B. 0.172
C. 0.12
D. 0.201
Define the following events:

B = Event of death
* P(B | B4)]
= (0.1 * 0.04)/[(0.1 * 0.04) + (0.29 * 0.05) + (0.49 * 0.1) + (0.12 * 0.14)
= 0.004/(0.004 + 0.0145 + 0.049 + 0.0168)
= 0.04745 or 4.7%
P(B2 | B) = (P(B2) * P(B | B2))/[ P(B2) * P(B | B2) + (P(B1) * P(B | B1) + (P(B3) * P(B | B3) + (P(B4) *
P(B | B4)]
= (0.29 * 0.05)/[(0.29 * 0.05) + (0.1 * 0.04) + (0.49 * 0.1) + (0.12 * 0.14)
= 0.0145/(0.0145 + 0.004 + 0.049 + 0.0168)
= 17.2%
* P(B | B4)]
= (0.49 * 0.10)/[0.49 * 0.10 + 0.1 * 0.04 + 0.29 * 0.05 + 0.12 *0.14]
= 0.049/(0.049 + 0.004 + 0.0145 + 0.0168)
= 58%
P(B4 | D) = 100 – 58 – 17.2 – 4.7 = 20.1% (Sum of all the probabilities must add up to 1).
51. An investment firm classifies capital projects into three different categories, depending on
risk level: Standard, Preferred, and Ultra-preferred. Of the firm’s projects, 60% are standard,
30% are preferred, and 10% are ultra-preferred. The probabilities of a project making a loss are
0.01, 0.005, and 0.001 for categories standard, preferred, and ultra-preferred respectively.
If a capital project makes a loss in the next year, then what is the probability that the project was
standard (correct to 2 decimal places)?
A. 0.79
B. 0.7895
C. 0.22
D. 0.15
Let:
L = Event a project makes a loss
S = Event of a standard project
P1 = Event of a preferred project
U = Event of a ultra-preferred project
We wish to determine P(S | L)

P(S | L) = (P(S) * P(L | S))/ [P(P1) * P(L | P1) + P(U) * P(L | U)]
= (0.6 * 0.01)/[(0.6 * 0.01) + (0.3 * 0.005) + (0.1 * 0.001)]
= 0.006/[0.006 + 0.0015 + 0.0001]
= 0.7895 or 79%
52. Upon arrival at a cancer treatment center, patients are categorized into one of four stages
namely: stage 1, stage 2, stage 3, and stage 4. In the past year,
i. 10% of patients arriving were in stage 1

ii. 40% of patients arriving were in stage 2
iii. 30% of patients arriving were in stage 3
iv. The rest of the patients were in stage 4
v. 10% of stage 1 patients died
vi. 20% of stage 2 patients died
vii. 30% of stage 3 patients died
viii. 50% of stage 4 patient died
Given that a patient survived, what is the probability that the patient was in stage 4 upon arrival?
(correct to 2 decimal places)
A. 0.13
B. 0.14
C. 0.138
D. 0.139
Let:
D = Event of death of a cancer patient
C1 = event of stage 1 cancer
We wish to determine P(C4 | D’) where D’ denotes survival

P(C4 | D’) = (P(C4) * P(D’ | C4))/[ (P(C4) * P(D’ | C4) + ((P(C1) * P(D’ | C1) + (P(C2) * P(D’ | C2) +
(P(C3) * P(D’ | C3)
= (0.2 * 0.5)/[(0.2 * 0.5) + (0.1 * 0.9) + (0.4 * 0.8) + (0.3 * 0.7)
= 0.1/(0.1 + 0.09 + 0.32 + 0.21)
= 14%
53. Upon arrival at a cancer treatment center, patients are categorized into one of four stages
namely: stage 1, stage 2, stage 3, and stage 4. In the past year,
i. 10% of patients arriving were in stage 1

ii. 40% of patients arriving were in stage 2
iii. 30% of patients arriving were in stage 3
iv. The rest of the patients were in stage 4
v. 10% of stage 1 patients died
vi. 20% of stage 2 patients died
vii. 30% of stage 3 patients died
viii. 50% of stage 4 patient died
Given that the patient died, what is the probability that the patient was in stage 4 cancer?
A. 0.86
B. 0.1
C. 0.36
D. 0.5
We wish to determine P(C4 | D)
P(C4 | D) = (P(C4) * P(D | C4))/[ (P(C4) * P(D | C4) + ((P(C1) * P(D | C1) + (P(C2) * P(D | C2) +
(P(C3) * P(D | C3)
= (0.2 * 0.5)/[(0.2 * 0.5) + (0.1 * 0.1) + (0.4 * 0.2) + (0.3 * 0.3)]
= 0.1/(0.1 + 0.01 + 0.08 + 0.09)
= 36%
54. You are an analyst at a large mutual fund. After examining historical data, you establish that
all fund managers fall into 2 categories: superstars (S) and ordinaries (O).
Superstars are by far the best managers. The probability that a superstar will beat the market in
any given year stands at 70%. Ordinaries, on the other hand, are just as likely to beat the
market as they are to underperform it. Regardless of the category in which a manager falls, the
probability of beating the market is independent from year to year. Superstars are rare
diamonds because only a meager 16% of all recruits turn out to be superstars.
During the analysis, you stumble upon the profile of a manager recruited 3 years ago, who has
since gone on to beat the market every year.
Determine the probability that the manager was a superstar when he was recruited into the
fund.
A. 0.5
B. 0.86
C. 0.7
D. 0.16
Let:
B = Event that a manager beats the market
S = Event that a superstar is recruited
Therefore,
P(B | S) = 70% = 7/10
P(B | O) = 50% = 1/2
At the time of recruitment, the probability of the manager being a superstar was just the
unconditional probability of a manager being a superstar i.e. P(S) = 16%
What is the probability that the manager is a superstar as at present?

A. 0.46
B. 0.34
C. 0.84
D. 0.16
We need to determine P(S | 3B): The probability that the manager is a superstar given that they
have managed to beat the market in three consecutive years. As such, we need to apply Bayes'
theorem.
P(S | 3B) = P(S) * P(3B | S)/P(3B)
Now, we already have P(S) = 16% = 4/25
P(3B | S) = (7/10)3 since performance is independent from one year to the next
= 343/1000
P(3B) = unconditional probability of beating the market in 3 consecutive years

= weighted average probability of 3 marketing-beating years over both superstars and
ordinaries
= P(3B | S) * P(S) + P(3B | O) * P(O)
= [(7/10) )3 * 4/25] + [(1/2) )3 * 21/25]
= (343/1000 * 4/25) + (1/8 * 21/25)
= 1372/25000 + 21/200
= 16%
Therefore,
16%*34.3%/16% = 34.3% or 0.343
What is the probability that the manager is NOT a superstar?

A. 0.66
B. 0.7
C. 0.45
D. 0.64
Since we already have the probability that the manager is a superstar now, the remaining
probability is surely the likelihood of the manager not being a superstar. i.e. being ordinary.
P(O | 3B) = 1 – P(S | 3B)
57. A human health organization tracked a group of individuals for 5 years. At the
commencement of the study, 25% were categorized as heavy smokers, 40% as light smokers
and the remaining as nonsmokers. Results revealed that light smokers were twice as likely as
nonsmokers to die during the half-decade study, but only half as likely as heavy smokers.
During the period, a randomly selected group member passed on.
Compute the probability that the individual who died was a heavy smoker.
A. 0.19
B. 0.53
C. 0.47
D. 0.175
Let:
D = Event of death
L = Event of light smoker
H = Event of heavy smoker
N = Event of nonsmoker
We need to calculate P(H | D)
Now, we already know that:

P(D | L) = 2P(D | N) and P(D | L) = 1/2P(D | H)
Applying Bayes’ theorem,

P(H | D) = (P(H) * P(D | H))/[(P(H) * P(D | H) + P(L) * P(D | L) + P(N) * P(D | N)]
= (2P(D | L) * 0.25)/[ (2P(D | L) * 0.25) + P(D | L) * 0.4 + 1/2P(D | L) * 0.35]
= 0.5/(0.5 + 0.175 + 0.4)
= 0.4651
58. Which of the following best describes the difference between the Bayesian approach and
the frequentist approach?
A. The Bayesian approach is based on a prior belief regarding the probability of occurrence of
an event
B. The frequentist approach is based on a prior belief regarding the probability of occurrence of
an event
C. The Bayesian approach is applicable with both small and large sample sizes while the
frequentist approach is only applicable with large sample sizes
D. While the Bayesian approach is based on a prior belief regarding the probability of
occurrence of an event, the frequentist approach involves drawing conclusions from sample
data using the frequency of the data
The Bayesian approach draws on a prior belief regarding the probability of an event occurring.
The frequentist approach, on the other hand, draws on the frequency of events occurring during
the most recent sample.
59. Peter selects a coin from a pair of coins and tosses it. While coin 1 is double-headed, coin 2
is a normal unbiased coin. After the toss, the result is a head. Calculate the probability that it
was coin 1 which was tossed.
A. 1/3
B. 2/3
C. 0.5
D. 0.75
We need to determine P(coin 1 | head). By applying Bayes’ theorem,

P(coin 1 | head) = P(coin 1) * P(head | coin 1)/[ P(coin 1) * P(head | coin 1) + P(coin 2) * P(head
| coin 2)
= (1 * 1/2)/[1 * 1/2 + 1/2 * 1/2 ]
= 2/3
Hypothesis Testing and Confidence Intervals

60. At a certain investment firm, each of the firm’s 5 managers is tasked with overseeing a
project. During a given one-year period, the managers reported the following individual returns
from their projects:
[24%, 26%, 30%, 18%, 20%]
Calculate the population variance of these returns.

A. 0.1824%
B. 18.24%
C. 22.8%
D. 0.228%
Note that the data given is comprised of the entire population and NOT a sample. As such, we
should use the formula for calculating the population variance.
We know that σ2 = [∑(Xi – μ)2]/N where N is the size of the population
And μ = (∑(Xi))/N = (0.24 + 0.26 + 0.30 +0.18 +0.20)/5 = 0.236
σ2 = [(0.24 – 0.236)2 + (0.26 – 0.236) 2 + (0.30 – 0.236) 2 + (0.18 – 0.236) 2 + (0.20 – 0.236)2]/5
= (0.000016 + 0.000576 + 0.004096 + 0.003136 + 0.001296)/5
= 0.001824 or 0.1824%
Note: Had we been given sample data, the formula for the mean would remain unchanged but
when calculating the variance, we would divide the sum of squared deviations by (n – 1) to
remove bias.
61. The mean hourly wage for coal workers in China is $15.5 with a population standard
deviation of $3.2. Calculate the standard error of the sample mean if the sample size is 30.
A. 3.2
B. 0.3413
C. 0.1067
D. 0.5842
Since the standard deviation for the population is known,

Standard error of the mean = σ/√n = 3.2/√30 = 0.5842
Interpretation: If we were to take a number of samples of size 30 from Chinese coal workers
population and proceed to prepare a sampling distribution of the sample means, the distribution
would have a mean of $15.5 and a standard error of $0.58.
Note: In most cases, σ is unknown in which case we work with the sample standard deviation, s.
62. A random sample of 50 FRM exam candidates was found to have an average IQ of 130.
The standard deviation among candidates is known (approximately 20). Assuming that IQs
follow a normal distribution, calculate a 2-sided 95% confidence interval for the mean IQ of FRM
candidates.
A. (125, 135)
B. 1.96 ± 5.5
C. (130, 135.5)
D. (124.5, 135.5)
For any sample that comes from a normally distributed population, we know that:
(x̄ - μ)/(σ/√n) ~ N(0,1) when σ is known
Thus, a 95% CI for the mean, μ = x̄ ± Zα/2 *σ/√n where α = 0.05

= 130 ± 1.96 * 20/√50
= 130 ± 5.5437
Which can be expressed as (124.5, 135.5).
63. An auto insurance company intends to establish the mean claim amount demanded by
policyholders who own SUVs. After extensive analysis of past records, the company believes
the standard deviation of such claims is about $200.
The company wishes to construct a 95% confidence interval for the mean claim amount such
that the interval is of width “±$50”. Determine the value of n, the sample size that would be
required to achieve this.
A. 62
B. 100
C. 30
D. 124
A 95% CI for the mean μ, assuming normal distribution = X ± 1.96σ/√n
Since σ = 200, we need to find the value of nsuch that:

1.96 * σ/√n = 50
1.96 * 200 = 50√n
392 = 50√n
√n = 7.84
n = 62
64. After 72 FRM Part 1 students wrote a mock exam, the mean score was 75. Assuming that
the population standard deviation is 10, construct a 99% confidence interval for the mean score
on the mock exam.
A. (75, 85)
B. (65, 75)
C. (71.96, 78.04)
D. (75, 78.04)
A 99% CI for the mean, μ = x̄ ± Zα/2 *σ/√n where α =0.01
From the normal dist. table, Z0.005 = 2.58
μ = 75 ± 2.58 * 10/√72
= 75 ± 3.04
Giving us a CI = 71.96 ≤ μ ≤ 78.04
65. Which of the following best defines the term “hypothesis” as used in statistics?
A. An assumption about a problem
B. The current state of knowledge or belief about the value of a population parameter
C. A test that divides the sample space into a region of acceptance and the critical region
D. A statement about the value of a population parameter developed with the intention of testing
a theory or belief
A hypothesis describes an assumption about the true value of a population parameter, such as
the mean. It is developed to test the validity and accuracy of a belief about the parameter. For
example, suppose a risk manager intends to establish the mean monthly return on a stock
option. They may formulate a hypothesis such as:
“The mean monthly return on stock options is positive” or;

“The mean monthly return on stock options is greater than 10%”
66. An investment firm intends to carry out a test to determine whether bonuses have any
significant effect on job performance. The head of the human resource department develops the
following sets of possible hypotheses.
I. H0: Bonuses do not have any effect on job performance

H1: Bonuses improve job performance
II. H0: Bonuses do not have any effect on job performance

H1: Bonuses reduce job performance
III. H0: Bonuses do not have any effect on job performance

H1: Bonuses affect job performance
IV. H0: Bonuses have no effect on job performance

H1: Bonuses improve job performance
Which of the above pairs of hypotheses implies a two-sided test?

A. I
B. II
C. III
D. IV
The difference between a one-sided test and a two-sided test is that while the alternative
hypothesis in former explores the possibility of a change in only one direction (increase or
decrease), the latter explores the possibility of a change in either direction. While the alternative
hypothesis in each of sets I, II, and IV explore increases or decreases, the word “affect” in the
H1 of set III leaves open the possibility of either an increase or a decrease in job performance.
67. Which of the following choices correctly defines type I error, type II error, and p-value?
Type I error Type II error P-value
I. The probability of The probability of The highest level at

rejecting H0 when it is accepting H0 when it which H0 can be
in fact true is in fact false rejected
II. The probability of The probability of The highest level at

accepting H0 when it rejecting H0 when it is which H0 can be
is in fact false in fact true rejected
III. The probability of The probability of The lowest level at

rejecting H0 when it is accepting H0 when it which H0 can be
in fact true is in fact false rejected
IV. The probability of The lowest level at The probability of

rejecting H0 when it is which H0 can be accepting H0 when it
in fact true rejected is in fact false
A. I
B. II
C. III
D. IV
Let’s borrow a little from the justice system: Suppose a judge sends a murder suspect to jail
when the suspect is in fact innocent; that would be type I error. On the other hand, suppose the
same judge frees the suspect when they are in fact guilty; that’s type II error. In both cases the
presumption (H0) would be “innocent”.
Similarly, let’s say we’ve got H0: μ = 0 and H1: μ ≠ 0. Rejecting H0 when μ is in fact equal to zero
constitutes a type I error. Failing to reject H0 when μ is in fact not equal to zero (either greater
than or less than zero) constitutes a type II error.
Lastly, the probability value (p-value) is the probability, assuming H0 is true, of observing a test
statistic at least as extreme as the value observed.
68. A random sample of 50 FRM exam candidates was found to have an average IQ of 125.
The standard deviation among candidates is known (approximately 20). Assuming that IQs
follow a normal distribution, carry out a statistical test (5% significance level) to determine
whether the average IQ of FRM candidates is greater than 120. Compute the test statistic and
give a conclusion.
A. Test statistic: 1.768; Reject H0
B. Test statistic: 2.828; Reject H0
C. Test statistic: 1.768; Fail to reject H0
D. Test statistic: 1.0606; Fail to reject H0
The first step: Formulate H0 and H1

H0: μ ≤ 120
H1:μ > 120
Note that this is a one-sided test because H1 explores a change in one direction only
Under H0, (x̄ - 120)/(σ/√n) ∿N(0,1)
Next, compute the test statistic:

= (125 – 120)/(20/√50) = 1.768
Please confirm that P(Z > 1.6449) = 0.05, which means our critical value is the upper 5% point
of the normal distribution i.e. 1.6449. Since 1.768 is greater than 1.6449, it lies in the rejection
region. As such, we have sufficient evidence to reject H0 and conclude that the average IQ of
FRM candidates is indeed greater than 120.
Alternatively, we could go the “p-value way”

P(Z > 1.768) = 1 – P(Z < 1.768) = 1 – 0.96147 = 0.03853 or 3.853%
This probability is less than 5% meaning that we have sufficient evidence against H0. This
approach leads to a similar conclusion.
69. A stock has an initial market price of $80. Exactly one year from now, its price will be given
by:
P = 80 * exp(i) where i is the rate of return.
i is normally distributed with mean 0.2 and standard deviation 0.3. Construct a 95% confidence
interval for the price of the stock after one year.
A. ($54.27, $80)
B. ($80, $175.91)
C. ($54.27, $175.91)
D. ($54.27, $140)
A 95% CI for the price = μ ± Zα/2 * σ where α =0.05

= μ ± 1.96 * 0.3
= μ ± 0.588
Therefore, P = 80 * exp(0.2 ± 0.588)

= ($54.27, $175.91)
70. A manager conducts a hypothesis test at the 1% significance level. What does this mean?
A. P(reject H0 | H0 is false) = 0.01
B. P(reject H0 | H0 is true) = 0.01
C. P(not reject H0 | H0 is false) = 0.01
D. P(not reject H0 | H0 is true) = 0.01
The significance level represents the probability of committing a type I error, or rejecting a
correct model. This is equivalent to P(reject H0 | H0 is true). Choice C is actually the probability
of committing type II error.
71. Suppose you conducted a hypothesis test. What would happen if you decrease the level of
significance of the test?
A. The likelihood of committing a type II error decreases
B. The likelihood of a type I error increases
C. The likelihood of rejecting the null hypothesis when it’s in fact true decreases
D. The likelihood of frequently committing a type I error increases, even when it’s in fact true
Having seen that the significance level gives the probability of rejecting a true null hypothesis, it
follows that a decrease in α (the level of significance) effectively decreases this probability. That
means a decrease of α from, say, 5% to 1%, would mean less frequent rejection of a true null
hypothesis.
72. Justin Heinz, FRM, suspects that the earnings of the insurance industry are more divergent
than those of the banking industry. In a bid to confirm his suspicion, Heinz collects data from a
total of 31 insurance companies and establishes that the standard deviation of earnings across
that industry is $4.8. Similarly, he collects data from 41 banks and establishes that the standard
deviation of earnings across that industry is $4.3. Conduct a hypothesis test at the 5% level of
significance to determine if the earnings of the insurance industry have a greater standard
deviation than those of the banking industry.
Choice Hypothesis Test Decision

statistic
I. H0: σ12 ≤ σ22 and 1.2461 Earnings are statistically not

H1: σ12 > σ22 significant from one another
II. H0: σ12 ≤ σ22 and 1.74 Earnings are statistically not
H1: σ12 > σ22 significant from one another
III. H0: σ12 ≤ σ22 and 1.2461 Earnings are statistically

H1: σ12 ≤ σ22 significant from one another
IV. H0: σ12 ≤ σ22 and 1.74 Earnings are statistically not
H1: σ12 < σ22 significant from one another
A. I
B. II
C. III
D. IV
As always, the first step involves formulating relevant hypothesis. We are concerned that the
earnings of the insurance could be greater (more variant) than those of the banking industry.
Therefore, the appropriate hypothesis is:
H0: σ12 ≤ σ 22 and H1: σ 12 > σ 22
Where σ 12 is the variance of earnings for the insurance industry, and σ 22 is the corresponding
variance for the banking industry.
Next in line is the selection of the test statistic. When comparing the variances of two different
populations, we use the F-statistic, computed as:
F = (S12/S22) where S12 and S22 are the sample variances
The F-statistic has (n1 – 1, n2 -2) degrees of freedom. i.e. F30,40
F = 4.82/4.32 = 1.2461
Note that this is a one-sided test. As such, our critical value should be the upper 5% point of the
F-distribution with (30, 40) degrees of freedom. This value = 1.74
Since 1.2461 is less than 1.74, it lies in the non-rejection and therefore we have insufficient
evidence to reject H0 at the 5% level of significance.
Decision: Heinz could argue that at the 5% level, the earnings of the insurance sector and those
of the banking sector are not significantly different from one another.
73. At 95% confidence interval, the value at risk (VaR) of a portfolio is approximately $10
million. During a 100-day period, the VaR was exceeded on 9 different occasions. Based on this
information:
A. This model is overestimating risk
B. This model is underestimating risk
C. This model is appropriate for estimating the risk
D. The model is accurate
The 95% CI means that we expect to have losses exceeding $10 million just 5% of the time.
Thus, during a 100-day period, we would expect a maximum of 5 exceedances. That 9 of these
were recorded means the model is underestimating risk, and is therefore NOT appropriate.
74. A population has a known mean of 100. Suppose 36 samples are randomly drawn from this
population with replacement. The observed mean is 97.8 and the standard deviation is 10.
Calculate the standard error of the sample mean.
A. 0.8165
B. 1.667
C. 1.8165
D. 16.67
The standard error of the sample mean = σ/√n = 10/√36 = 1.6667
Linear Regression with One Regressor
75. Which of the following is NOT true regarding a scatter plot?

A. It’s a visual representation of the relationship between the explained variable and the
independent/explanatory variable
B. It’s a standard two-dimensional graph where the values of the explained variable are plotted
on the X-axis while those of the explanatory variable are plotted on the Y-axis
C. It reveals the kind of relationship between the explained variable and the explanatory variable
– positive, negative, or linear.
It’s imperative to note that the independent/explanatory variable takes on the X-axis while the
explained/dependent variable takes on the Y-axis. For example, when assessing the
relationship between productivity and the age of employees, age would be our independent
variable (X-axis) while productivity would be the dependent variable (Y-axis).
76. The relationship between two variables can be explained by the following regression
function:
Yi = B0 + B1 × Xi + εi
What does εi represent?

A. The difference between total variation and the explained variation
B. The effects of independent variables not included in the model
C. The slope coefficient
D. The intercept coefficient
εi denotes the effects of independent variables other than the variable of interest to the
researcher, which nonetheless impact the dependent variable. For example, assume you want
to assess the effect of class size (independent variable) on student performance (dependent
variable): the error term might comprise of factors such as teacher quality, student economic
background, or even luck.
77. What is the difference between regression analysis and correlation analysis?
A. Regression enables us to measure the association between two or more variables in
specified units of the dependent variable
B. Regression enables us to establish a line of best fit through the data
C. With regression, we can predict the values of the dependent variable
D. All of the above
While correlation merely informs us of the existence of a linear relationship, regression analysis
goes further to enable us to predict distinctive values of the dependent variable given distinctive
values of the independent variable. The regression line is very useful in estimation and
prediction.
78. A hospital uses ultrasound technology to measure the weight of unborn babies as follows:
Gestation period in weeks 30 32 34 36 38 40
Estimated weight of foetus 1.6 1.7 2.5 2.8 3.2 3.5
Further information: SXX = 70, SYY = 3.015, SXY = 14.3
Calculate the least square estimators of the slope and the Y-intercept (in that order).
A. 0.2043, -4.6
B. 0.20, -4
C. 2.55, 35
D. 0.2043, 35
Under OLS estimation,

The regression equation is of the form: yi = α + βxi
Where Y is the dependent variable (foetal weight), X is the independent variable (gestation
period), α = the y-intercept, and β = the slope.
β = ∑((xi - x̄)(yi - y-bar)/( xi - x̄)2 = Sxy/Sxx = 14.3/70 = 0.2043
α = y-bar - βx̄
y-bar = (1.6 + 1.7 + 2.5 + 2.8 + 3.2 + 3.5)/6 = 2.55
x̄ = (30 + 32 + 34 + 36 + 38 + 40)/6 = 35
Therefore, α = 2.55 – 0.2043 * 35 = -4.60
79. Given a regression equation is of the form: yi = α + βxi where Y is the dependent variable
(foetal weight), X is the independent variable (gestation period, in weeks), α = the y-intercept,
and β = the slope.
If α = -4.60 and β = 0.2043, then estimate the weight of the foetus at exactly 31 weeks.
A. 1.533kg
B. 1.733kg
C. 1.722kg
D. 1.8kg
The regression equation is of the form: yi = α + βxi
Therefore, weight = -4.6 + 0.2043*31 = 1.733kgs
Quantifying Volatility in VaR Models

80.Linda Christian is a risk consultant at Garson Partners. She is one of an influential and
respected risk manager in the risk management community of Wall Street banks. In a recent
publication, she shared her thoughts on the use of firm-wide VaR at major banks. From the
following, determine which option is least consistent with the actual features of VaR in banks?
A. Firm-wide VaR does not include the funding liquidity risk of banks
B. Firm-wide VaR is not likely to account for all the risks, especially credit risk
C. The higher the correlation between market risk and credit risk, the higher the firm-wide
VaR will be
D. Due to a lack of historical data, it is difficult to determine if the daily VaR is unbiased or
biased
Credit risk is included in VaR measurements, but the operational risk is not, which is why Firm-
wide VaR does not account for all of the risks. Option A is accurate because the firm-wide VaR
measure does not include the funding liquidity risk of the bank, which refers to a sudden
decrease in the funding. Option C is accurate because a high the correlation between credit risk
and market risk will result in higher VaR levels. Option D is accurate because a lack of historical
data makes it difficult to determine if the daily VaR is unbiased or biased upwards or
downwards.
81. Billy Marquette has recently joined a small company that provides private commercial jets to
royal families, government officials, and directors of big firms. Marquette is a retired commercial
pilot with a very basic understanding of finance. On his first day, he is handed a report on risk
management measures. The excerpt from the report says “due to volatility in oil prices, the
company has weekly 10% VaR as €20,000”. Which of the following is the most appropriate
explanation of the excerpt?
A. It indicates that there is a probability that the company can experience a loss of €2,000 on a
weekly basis
B. It indicates that there is a probability that the company can experience a loss of 10% or
€20,000 in any given week
C. It indicates that there is a 10% probability in any given week that the company can
experience a loss €20,000 or more
D. It indicates that there is a 10% probability in any given week that the company can
experience a loss up to €20,000
The “weekly 10% VaR as €20,000” means that there is a chance or probability of 10% in any
given week of its operation that the company can experience a loss of €20,000 or more. VaR is
a probabilistic risk measure that measures the potential loss in the value of the portfolio at any
given time.
82. A number of risk measures are based on the parametric approach, which assumes that the
asset returns are normally distributed. However, mathematician and statisticians have
discovered that in reality, the asset returns deviate from normality. Which of the following
options is least likely consistent with the assumption that the asset returns deviate from
normality?
A. Asset returns have fat-tailed distributions which means assets have a higher probability
weight in their tails relative to the normal distribution
B. Asset returns have skewed distribution, which means that the declines in asset prices are
more severe than increases in prices
C. Asset returns have unstable parameter values due to varying market conditions
D. Asset returns have symmetrical distributions which means they are evenly distributed around
the mean returns
Option D is least consistent with the assumption that the asset returns deviate from a normal
distribution. Symmetric distribution is the assumption of the normal distribution that indicates
that the returns are evenly distributed around it means that also indicates that there is no
skewness in asset returns.
83. Donald York is a quantitative analyst at Brooklyn Investments Hub, a tech investment
company based in New York. York brings 5 years of experience in quantitative and statistical
analysis. In one of his explanatory article, he mentioned that when the mean and standard
deviation of asset returns are the same for any given time period, the distribution of returns is
said to be an unconditional distribution. In contrast, if the mean and variance of the return
change over time, the return distribution is referred to as a conditional distribution. Identify the
correct option from the following.
A. York’s explanation regarding the unconditional distribution is incorrect
B. York’s explanation regarding the conditional distribution is incorrect
C. York’s explanation regarding the conditional and unconditional distribution is incorrect
D. York’s explanation regarding the conditional and unconditional distribution is correct
Both definitions are correct. An unconditional distribution is referred to as a case when the mean
and the standard deviation of the asset returns are the same for any given time periods. On the
other hand, due to changing economic and market dynamics, the mean and variance of asset
returns can change over time. The returns distributions with changing means and variances are
referred to as conditional distributions.
84. An analyst is conducting a fundamental and technical analysis on Pak-China Trading Co

(PCTC) stocks. The analysis takes into account the stock’s daily returns based on the mean and
volatility of returns. If PCTC returns have a mean of 9.56 bps/day, and a high volatility of 14.25
bps/day and a low volatility of 6.18 bps/day, then determine which of the following distributions
fits the most the characteristics of PCTC’s return distribution?
A. Unconditional distribution
B. Regime-switching distribution
C. Unconditionally lognormal distribution
D. Conditional distribution
In a regime switching distribution, the volatility of returns switches between low volatilities to
high volatilities. In this distribution, the volatility does not switch to any volatility between the high
and low volatility regimes.
85. VaR is the widely used and reliable measure of assessing risk. One of the most common
ways of estimating VaR is through analyzing historical time series data and deriving the shape
of the distribution of risk. Which of the following VaR estimating approaches does not use
historical data?
A. Implied volatility approach
B. Hybrid approach
C. Parametric approach
D. Nonparametric approach
The implied volatility approach does not use historical time series data to determine the shape
of the distribution of risk. Rather, it uses derivatives pricing models such as the Black-Scholes
options pricing model to derive the distribution using implied volatility. On the other hand, the
parametric approach, nonparametric approach, and hybrid approach (mix of parametric and
non-parametric) use historical data.
86. Bloomsdale Investment uses a number of models to assess the VaR of asset returns under
its portfolio. Most of the models it uses belong to a distinctive category of VaR assessing model
classes that use historical data. Which of the following model classes imposes specific
distributional assumptions on conditional asset returns while assessing VaR?
A. Hybrid approach
B. Parametric approach
C. Nonparametric approach
D. Implied volatility approach
The category of parametric approach models imposes specific distributional assumptions on

conditional asset returns. An example of a model under this class of risk assessing models is
the conditional (log) normal model with time-varying volatility.
87. In which of the following categories of VaR estimating models can you accurately categorize
the Black-Scholes model?
A. Hybrid models class
B. Parametric models class
C. Nonparametric models class
D. Implied volatility models class
he correct answer is: D).
The implied volatility models class uses derivatives pricing models like the Black-Scholes
options pricing model to estimate VaR using implied volatility. Other classes of models such as
the parametric, nonparametric and hybrid (mix of parametric and non-parametric) classes use
historical data to determine the shape of the distribution of risk.
88. Xaomi Lee, a fresh graduate from the Boston University of Massachusets, has recently
joined a large investment bank as a trainee in the risk management unit. In the second month of
her joining, her senior manager asked Lee to use a VaR estimate model for a risk analysis on a
specific asset. Which of the following VaR estimating models is the simplest model to use for
the trainee?
A. Parametric models
B. Nonparametric models
C. Implied volatility models
D. Hybrid models
The nonparametric model is the simplest model for estimating VaR. As compared to the
parametric model, the nonparametric model does not impose a specific set of distributional
assumptions; rather it uses historical data directly.
89. Gareth Graham is a senior risk consultant for Poincare Consulting Group. Graham has a
strong reputation in the risk managers community, which is why he is frequently invited as a
guest speaker at various business schools. During a recent seminar at a reputable business
school in Vancouver, Graham mentioned the following comments regarding the cyclicality of
volatility:
I. It should be considered while analyzing the risk of financial assets that volatility in financial
markets is time-varying
II. While using a historical data model for analyzing volatility, more weight should be put into
recent data as opposed to earlier data
Which of the following is correct?

A. Only statement I is correct
B. Only statement II is correct
C. Both statements are correct
D. None of the statements are correct
Both statements are correct. It should be taken into account while analyzing the risk of assets
returns that the volatility in the financial markets is time-varying and sticky. It should also be
assumed that the more recent data of asset returns provides more information about future
volatility.
90. An analyst is comparing the STDEV or GARCH methodology with that of the RiskMetric®
approach for estimating VaR using historical data. He wrote down the following similarities
between both methods. Which of the following similarities is incorrect?
A. Both methods belong to the parametric class of risk assessing models
B. Both methods attempt to estimate conditional volatility
C. Both methods apply equal weights to all the periods
D. Both methods use recent historic data for assessing risk
The standard deviation models STDEV or GARCH apply equal weights to all the windows of
past data, while the RiskMetric approach applies higher weights on more recent data. The
weights decline exponentially to zero as returns become older.
91. Haresh Kumar, a fresh graduate from an Ivy League institute, has recently applied for a risk
analyst position in an investment bank’s risk management unit. During the interview, a senior
manager asked Kumar to identify some of the advantages of nonparametric methods of
estimating VaR. He provided the following four advantages. Identify the one that is incorrect.
A. For the estimation of VaR, nonparametric models do not require assumptions related to the
distribution of returns
B. In nonparametric models, fat tails, skewness, and other deviations from distribution are not
considered in the estimation process of VaR
C. The multivariate density estimation (MDE), a nonparametric model, allows varying weights
based on the relevance of data, rather than the timing of the data
D. The RiskMetrics® model is a nonparametric approach that exponentially smoothens weights
on historical data
RiskMetrics® is a parametric model. Options A, B and C are appropriate advantages of using

nonparametric model for estimating VaR.
92. Which of the following is not a disadvantage of the nonparametric approach of estimating
VaR as compared to the parametric approach?
A. Since the data is used less efficiently in nonparametric methods as compared to parametric
methods, large sample sizes are required to estimate the volatility accurately
B. Separating the full sample of data into different market regimes reduces the amount of usable
data for historical simulation models in the nonparametric approach
C. In nonparametric models, fat tails, skewness, and deviations from distribution are ignored in
the estimation process of VaR
D. The multivariate density estimation (MDE), a nonparametric model, requires a large amount
of data that is directly related to the number of conditioning variables used in the model
“In nonparametric models, fat tails, skewness, and deviations from distribution are ignored in the
estimation process of VaR” is not a disadvantage, but an advantage of using a nonparametric
method for estimating VaR. Options A, B, and D are disadvantages of using nonparametric
methods as compared to parametric methods of estimating VaR.
93. Selma Kaya is a junior risk analyst at Galileo Investment Bank. She is interested in
estimated the joint density of returns of Algo Corp. and economic growth. Which of the following
models should Kaya most likely use?
A. Risk Metric®
B. Black-Scholes model
C. GARCH
D. Multivariate density estimation (MDE) model
The Multivariate density estimation (MDE) model allows estimating the joint density of returns of
assets and other specified variables like the slope term structure, CPI, economic growth, etc.
94. Samuel Juan has recently joined Arka Investments. Previously, Juan worked as a risk
analyst at one of the well-known credit rating agencies where he learned various advanced
methodologies for assessing risk through estimating VaR. At Arka, Juan is estimating the 10%
VaR of a specific asset by using historical simulations to estimate the percentile of returns and
using exponentially declining weights on past data. This methodology is most likely an example
of:
A. Parametric VaR estimating
B. Nonparametric VaR estimating
C. Hybrid VaR estimating
D. Implied Volatility VaR estimating
The hybrid VaR estimating methodology is a combination of parametric and nonparametric

methods. Juan’s method of estimating VaR by using historical simulations (nonparametric
method) to estimate the percentile of the returns and using exponentially declining weights
(parametric method) on past data is an example of the hybrid methodology.
95. Markus Schmidt is an independent risk consultant at a well-known audit firm. He is currently
working as an external risk advisory for George Reed Shipping Inc. During a meeting with the
senior management of the shipping company, Schmidt made the following comments:
I. Using nonparametric is a simple process as it does not impose a specific set of distributional
assumptions; rather it uses the historical data directly
II. Nonparametric methods are better predictors of the future volatility
Which of his comments is/are incorrect?

A. Comment I only
B. Comment II only
C. Both comments
D. None of the comments
The most efficient predictors of future volatility are not nonparametric models but implied
volatility models. Implied volatility models such as the Black-Scholes option pricing model can
react better and quicker to current market conditions.
96. In a board meeting at a large Frankfurt bank, one of the senior risk managers provided the
following disadvantages of using implied volatility methods for estimating VaR:
I. As the volatility in implied volatility methods is model dependent, a misspecified model can
generate faulty forecasts
II. Empirical results suggest that the implied volatility is biased downward, as it is generally
lower than realized volatility
Which of these disadvantages is/are incorrect?

A. Only point I
B. Only point II
C. Both points
D. None of the points
The point regarding the downward bias of implied volatility is incorrect. Empirical results have
shown that the implied volatility does demonstrate an upward bias as it is, on average, higher
than the realized volatility. Point I is correct because the volatility in implied volatility methods is
model dependent and can result in a wrong forecast if the model is misspecified.
97. The process of assessing the risk of assets returns does not end at estimating VaR using
the appropriate methodology, but the risk assessing also requires the testing of VaR. The
process of backtesting is used to compare losses as estimated by the VaR to those actually
incurred over the given testing period. Which of the following attributes of VaR is least likely
examined through the backtesting process?
A. The unbiasedness of VaR estimations
B. The adaptability of VaR estimations
C. The correctness of historical data used for VaR estimations
D. The robustness of VaR estimations
Backtesting does not evaluate the correctness and accuracy of the historical data used in the
estimation process of VaR. However, the backtesting method of comparing the losses estimated
by the VaR to those actually incurred over the given testing period is essential for evaluating the
unbiasedness, robustness, and adaptability of VaR estimations.
98. If the covariance between Japanese and English interest rates is 0.089, and the variances
of interest rates in Japan and England are 17.64% and 10.24%, respectively, then which of the
following is closest to the correlation between Japanese and English interest rates?
A. 4.927
B. 1.028
C. 0.6622
D. 0.8736
Correlation = Covariance / (Standard deviation (A) * Standard deviation (B))

Correlation = 0.089 / (√0.1764 * √0.1024) = 0.6622
99. Arthur Bell is the portfolio manager at FFF Investments. Recently, he bought 5,000 call
options on stocks of one of the local growth-oriented oil refining companies that have never paid
dividends. The strike price of the options was $50. The underlying stock is trading at $58 and
has the annual volatility of return of 33%. Bell estimated the delta of these options to be 0.55.
What is the approximate weekly 99% VaR of the position assuming 252 trading days and 52
trading weeks in the year?
A. $7,725.57
B. $17,007.01
C. $6,659.97
D. $14,661.22
VaR of the underlying = (33%)/√52 * 2,33 * $58 = $6.18
VaR of the position = 0.55 * $6.18 * 5000 = $17,007.01
100. The following are key assumptions of OLS for estimation of parameters. Which one is
NOT?
A. There is a random sampling of observations.
B. The expected value of the error term, conditional on the independent variable, is zero.
C. There are no large outliers in the observed data.
D. A linear relationship exists between the independent and the dependent variables.
Although choice D is also a relevant assumption, it’s not a key assumption. The key
assumptions have a lot to do with the error term and the distribution of observed values.

FRM Test 12 - Topic - Book 2 - 4

Uploaded by

Copyright:

Available Formats

FRM Test 12 - Topic - Book 2 - 4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

FRM Test 12 - Topic - Book 2 - 4

Uploaded by

Copyright:

Available Formats

FRM Part – 1 – Test ID – 0012 – Topic - Book 2 & 4 – Questions - 100

1. Which of the following can be categorized as continuous random variables?

The correct answer is: B).

2. Consider the following probability function for a discrete random variable X:

X = {10, 20, 30, Y,}, P(x) = x/100, otherwise P(x) = 0

Find the value of Y.

f(x) = x/5000 for 0 ≤ x ≤ 100

X = {10, 20, 30, 40}, P(x) = x/100

The multiplication rule of probability states that:

6. Two events are said to be independent and mutually exclusive if:

We can easily establish that:

We need to determine P(A ᴖ M ᴖ S):

P(M ᴗ S) = P(M) + P(S) – P(M ᴖ S),

P(M) = 0.6 + 0.15 – 0.45 = 0.3

Similarly, P(A ᴗ S) = P(A) + P(S) - P(A ᴖ S)

Therefore applying equation (I),

Which gives us P(A ᴖ M ᴖ S) = 0.1

From first principles, it can be shown that:

As such, P(A | B) = P(A ᴖ B)/P(B)

If A and B were independent, then P(A ᴖ B) = P(A) * P(B)

Normal 30% 20%

All the joint probabilities must add up to 1. Therefore,

P(R < -5%) = 16%

The formula for calculation of sample variance is s2 = 1/(n – 1)[∑x2 – n x̄2]

From the data, x̄ = 31353/100 = 313.53

s2 = 1/99[10,687,041 – 100 * 313.532] = 8654.64

The sample standard deviation = √s2 = √8654.64 = 93

Making ∑x2 the subject of the formula,

Removing the fraudulent claim gives,

Now, using the data for the remaining 8 claims,

Therefore, s = √12,145.2 = 110.2

Probability EPS Interest rates Beta

20% $1.5 21 1.15

10% $2.0 20% 10%

30% $1.3 20% 1.25

40% $1.2 10%

Compute the expected earnings per share.

EPS = E(X) = ∑P(xi)xi

State Probability(state) Return for stock A Return for stock B

Normal 35% 10% 15%

Slow 25% 8% 12%

The correct answer is: D).

Cov(A, B) = ∑P(s) * [RA – E(RA)] * [RB – E(RB)]

State P(S) RA RB P(S) * [RA – E(RA)] * [RB

The correct answer is: C).

20. A discrete random variable Y has probability function given by:

P(Y = y) 0.3 0.6 0.1

The variance of any given random variable is given by:

E(Y) = ∑YP(Y =y) = 0 * 0.3 + 1 * 0.6 + 2 * 0.1 = 0.8

E(Y2) = ∑Y2P(Y =y) = 02 * 0.3 + 12 * 0.6 + 22 * 0.1 = 1

Therefore, Var(Y) = 1 - 0.82 = 0.36

Let F be the proportion of females.

Applying the idea of a weighted mean,

Cov[E,F] = Cov[X + Y,X – Y]

P(Y = y) 0.3 0.2 0.4 0.1

Determine the variance of X, where X = 2Y + 10

V[X] = V[2Y + 10]

V[Y] = E[Y2] – E2Y

= 0.3 * 02 + 0.2 * 12 + 0.4 * 22 + 0.1 * 32 – [0.3 * 0 + 0.2 * 1 + 0.4 * 2 + 0.1 * 3] 2