FRM Part 1

Book 2 - Foundations of Risk Management

Chapter 3

DISTRIBUTIONS
 Learning Objectives
After completing this reading you should be
able to:
 Distinguish the key properties among the following distributions: uniform
distribution, Bernoulli distribution, Binomial distribution, Poisson
distribution, normal distribution, lognormal distribution, Chi-
squared distribution, Student’s t, and F-distributions, and identify
common occurrences of each distribution.
 Describe the central limit theorem and the implications it has when
combining independent and identically distributed (i.i.d.) random
variables.
 Describe i.i.d. random variables and the implications of
the i.i.d. assumption when combining random variables.
 Describe a mixture distribution and explain the creation and
characteristics of mixture distributions.
 Uniform Distribution
 A Uniform distribution has equally likely values over the range of
possible outcomes, say 𝛼 to 𝛽.
 X ~ U(𝛼,𝛽) is often written as shorthand for “the random variable X
has a continuous uniform distribution between 𝛼 and 𝛽.”

0 𝛼 𝛽 x
 All values of x from 𝛼 to 𝛽 are equally likely in the sense that the
probability that x lies in an interval of width Δx entirely contained in
the interval from 𝛼 to 𝛽 is equal to Δx/(𝛽 - 𝛼), regardless of the exact
location of the interval.
 Uniform Distribution
 The probability density function can be represented as:
𝟏
𝒇𝑿 𝒙 = 𝜶<X<𝜷
𝜷−𝜶
𝜶+𝜷 𝜷−𝜶 𝟐
 Mean = 𝝁 = Variance =
𝟐 𝟏𝟐
 Examples of uniform distributions include:
o A random number generator (say, between 0 and 1).
o The amount of time, in minutes, that a client spends waiting for
their turn to see a portfolio manager in an office may be uniformly
distributed between zero and 15 minutes, inclusively.
 Bernoulli Distribution
 A Bernoulli trial is an experiment which has (or can be regarded as
having) only two possible outcomes: 1 (“success”) and 0
(“failure”).

 The words “success” and “failure” are merely labels – they do not
necessarily carry with them the ordinary meanings of the words.
𝐏[𝐗 = 𝟏] = 𝐩 and 𝐏[𝐗 = 𝟎] = 𝟏 − 𝐩
 For example in the illustrative figure above, p = 0.7
 Bernoulli Distribution
 The mean and variance of the distribution are computed as:
𝐌𝐞𝐚𝐧 = 𝛍 = 𝐩 × 𝟏 + (𝟏 − 𝐩) × 𝟎 = 𝐩
𝐕𝐚𝐫𝐢𝐚𝐧𝐜𝐞 = 𝝈𝟐 = 𝐩 × 𝟏 − 𝒑 𝟐 + (𝟏 − 𝐩) × 𝟎 − 𝒑 𝟐 = 𝐩(𝟏 − 𝐩)

Examples
 In life insurance, “success” could mean “death” and failure could
mean “survival.”
 In any given year, a company may pay a dividend on common stock
(success) or pay nothing at all (failure).
o Suppose both events are equally likely, then:
I. Dividend = Success = P = P(X = 1) = 0.5
II. No dividend = Failure = 1 – p = P(X = 0) = 0.5
 Binomial Distribution
 Consider a sequence of n Bernoulli trials as above such that:
o The trials are independent of one another, i.e. the outcome of
any trial does not depend on the outcomes of any other trials; and
o The trials are identical, i.e. at each trial P(success) = p.
 Such a sequence is called a “sequence of n independent, identical,
Bernoulli (p) trials.”
 If X is the number of successes that occur in n trials:
𝐏 𝐗 = 𝒙 = 𝒏 𝑪𝒙 𝒑𝒙 𝟏 − 𝒑 𝒏−𝒙
Where n = 0, 1, 2… n; 0 < P < 1

n!
Recall that = n Cx =
n−x !x!

 If X is distributed binomially with parameters n and p, then we can

write,
X ~ Bin(n, p)
 Key Properties of the Binomial
Distribution
Moments
 μ = np
 σ2 = np 1 − p
o Very often, when using the binomial distribution, p will be written as 𝛉,
the probability of success. Rather than using 1 – p for failure, we would
use 1 − θ.
Example
 What is the probability that at least 9 out of a group of 10 banks who have
been hit by a serious cash crunch will ride out the crunch if the probability
of a bank surviving a cash crunch is 70%? (p = 0.7)
Solution
 The number of survivors is distributed binomially with parameters n = 10,
and p = 0.7 . If X is the number of survivors, then:
o P X ≥ 9 = P X = 9 + P X = 10
o = 10 C9 0.79 0.3 10−9 + 10 C10 0.710 0.3 10−10 = 0.1493
 Poisson Distribution
 This distribution models the number of events that occur in a specified
interval of time, when the events occur one after another in time in a well-
defined manner.

 A Poisson Process is a model:

 Series of discrete events (events occur singly)
 The average time between events is known (constant rate)
 The timing of events is random (non overlapping time intervals are
independent)

 The average number of events that occur within a specified period of time is
known, denoted as lambda (𝜆).

 Distribution:
𝒆−𝝀 𝝀𝒙
P(X = x) = ,
𝒙!
Where x = 0, 1, 2, 3, …,; 𝜆 > 0
 Poisson Distribution
If X has a Poisson
distribution with
parameter 𝜆, then we can
write X ~ Poi(𝜆).

Crucially, if X follows the

Poisson distribution, then • The number of clients arriving at a
bank per unit of time, say, an hour.
its mean and variance
• The number of goals scored by a
are both equal to 𝜆. soccer team in a match.
• The number of claims that an
insurance company receives per unit of
Examples of events that time.
follow a Poisson • The number of road crashes in a
distribution may include: country per unit of time.
 Poisson Distribution
Example
 The number of new clients a wealth management company receives
in a month is distributed as a Poisson random variable with mean
2.
 Calculate the probability that the company receives exactly 30
clients in a year.
Solution
 The number of clients in a year (2 × 12) has a Poi(24) distribution.
𝑒 −24 2430
 𝑃 𝑋 = 30 = = 0.0363 or 3.63%
30!
 Normal Distribution
 Most important probability distribution in statistics particularly
because it fits many natural phenomena.
o For example, salaries, test scores, blood pressure, and IQ scores
follow the normal distribution.
 Also known as the Gaussian distribution and the bell curve.
 Has 2 parameters: mean 𝝁 and the standard deviation 𝝈.
o And is symmetrical about 𝜇.
Equal halves

𝝈 𝝁 𝝈
 Normal Distribution
 The notation used for the Normal distribution is X ~ N(𝝁,𝝈𝟐 ).
 The following is the formula of a PDF that is normally distributed,
for a given random variable X:
𝟏 𝟏 𝒙−𝒖 𝟐
−𝟐 𝝈
𝒇 𝒙 = 𝒆
𝝈 𝟐𝝅
 It is not possible to find an explicit expression for Fx (x) = P(X ≤ x) ,
so tables have to be used.
𝐱−𝛍
o These are provided for the distribution of Z = , which is the
𝛔
standard normal variable.
o Z has mean 0 and standard deviation of 1.

To find probabilities using the normal distribution, the variable has to be

standardized (converted into a z-value).
 Normal Distribution
 The z-value measures how many standard deviations the
corresponding x value is above or below the mean.

x-value z-value
𝜇 0
𝜇 + 𝟏𝜎 1
𝜇 + 𝟐𝜎 2
𝜇 + 𝒏𝜎 n

68–95–99.7 rule
 Normal Distribution

Example: P(Z < 0.45) =

0.6736

 The probabilities in the table are “left hand” probabilities; they give
P(Z < z).
 The numbers in the first column are z-values that have only one
decimal place. The columns to the right supply probabilities for z-values
with two decimal places.
 For more probabilities more precise than the ones tabulated,
interpolation is necessary.
 Lognormal Distribution
 A continuous random variable X follows a lognormal distribution if its
natural logarithm, ln(X), follows a normal distribution.
o We can also say that if the natural log of a random variable, ln(X), follows
a normal distribution, the random variable, X, follows a lognormal
distribution.
 The lognormal distribution is an asymmetric distribution with interesting
applications for modeling the probability distributions of stock prices and
prices of other assets.
 It’s probability density function is:
𝟏 𝒍𝒏 𝒙 − 𝝁 𝟐
𝟏 −𝟐 ( )
𝒇 𝒙 = 𝒆 𝝈
𝒙 𝝈 𝟐𝝅

𝟏
𝛍+𝟐𝛔𝟐
 Mean = E [X] = 𝐞
𝟐 𝟐
 Variance = E[X− E [ 𝑿 ]𝟐 ) = 𝒆𝛔 − 𝟏 𝒆𝟐𝛍+𝛔
 Lognormal Distribution

 The lognormal distribution:

o Is asymmetric (skewed to the
right).
o Is bounded below by 0 (lowest
possible value).
o Fits well data on asset prices
(note that prices are bounded
below by 0).
o Note: The normal distribution fits
well data on asset returns.
0
 Student’s t-distribution
 Student’s t-distribution, or the t-distribution, is a bell-shaped
probability distribution, symmetrical about its mean.
 The t-distribution is the appropriate distribution to use:
i. When testing hypotheses and constructing confidence intervals
based on small samples (n < 30).
ii. When the underlying population is approximately normal but its
variance is unknown.

A t-distribution is like a Z distribution, except has slightly fatter

tails to reflect the uncertainty added by estimating . The bigger
the sample size (i.e., the bigger the sample size used to estimate
), then the closer t becomes to Z.
 Student’s t-distribution
• Student’s t-distribution, or the t-distribution, is a bell-shaped probability
distribution, symmetrical about its mean.
Note: t Z as n increases

Standard Normal
(t with df = )

t (df = 12)
t-distributions are bell-shaped and
symmetric, but have ‘fatter’ tails
than the normal distribution t (df = 6)

t
0
 Chi-square distribution
 The chi-square distribution results when ν independent variables
with standard normal distributions are squared and summed.
 It is used in hypothesis tests and in determining confidence intervals.
 Two common examples are:
o The chi-square test for independence in contingency tables; and
o The chi-square test to determine if the variance (or standard deviation) of
a population is equal to a pre-specified value.
 The chi-square distribution is asymmetrical, bounded below by zero, and
approaches the normal distribution in shape as the degrees of freedom
increase.
 Chi-square distribution
P

𝝌𝟐
 The chi-square test statistic, 𝜒 2 , with n – 1 degrees of freedom is
computed as:
𝟐 𝒏 − 𝟏 𝒔𝟐
𝝌𝒏−𝟏 =
𝝈𝟐𝟎
o Where n = sample size;
o s2 = sample variance
o 𝜎02 = hypothesized value for the population variance.
 F-distribution
 If two independent random variables, X and Y, have 𝜒 2 distributions with
parameter n1 and n2 respectively, then the function:
𝑿/𝒏𝟏
𝒀/𝒏𝟐
is said to have an F distribution with parameters “degrees of freedom” n1
and n2.
 The F distribution is an asymmetric distribution that has a minimum value
of 0, but no maximum value.
 Notably, the curve approaches but never quite touches the horizontal axis.

F The F-distribution is often used in

the analysis of variance (ANOVA)

0
 Mixture Distribution
 A mixture distribution is a mixture of two or more probability
distributions.
 Its formation may be necessary when the data at hand does not
match any particular distribution in which case a newly created
distribution may assist with explaining the relevant data.
 Central Limit Theorem (CLT)
 If we have i.i.d. random variables, X1, X2, ... , Xn, each with mean μ and
standard deviation 𝝈, and we define Sn as the sum of those n variables, then:
 𝐥𝐢𝐦 𝑺𝒏 𝒏→∞ ~ 𝑵(𝒏𝝁, 𝒏𝝈𝟐 )

In essence, ANY distribution tends toward the normal distribution as

n increases. (The larger the sample size, the better will be the normal
approximation to the sampling distribution of x).

 In finance, the CLT is applicable when a normal distribution is used to justify the
approximation of variables that have a financial aspect.
How large is large?
 The CLT applies when sample size is greater or equal than 30. In most
applications with financial data, sample sizes will be significantly greater than
30.
 Book 2 – Quantitative Analysis
Chapter 3
DISTRIBUTIONS
Learning Objectives Recap:
 Distinguish the key properties among the following distributions: uniform distribution,
Bernoulli distribution, Binomial distribution, Poisson distribution, normal distribution,
lognormal distribution, Chi-squared distribution, Student’s t, and F-distributions, and
identify common occurrences of each distribution.
 Describe the central limit theorem and the implications it has when combining
independent and identically distributed (i.i.d.) random variables.
 Describe i.i.d. random variables and the implications of the i.i.d. assumption when
combining random variables.
 Describe a mixture distribution and explain the creation and characteristics of mixture
distributions.

N EXT
BAYESIAN ANALYSIS