Professional Documents
Culture Documents
FRM-I - Book Quant Analysis
FRM-I - Book Quant Analysis
FRM-I - Book Quant Analysis
Quantitative Analysis - I
EduPristine www.edupristine.com
Agenda
Probability
Counting Principle
Random variable
Probability Concepts, Bayes Theorem
Basic Statistics
Mean, Median, Mode, Variance, Covariance, Correlation
Skewness, Kurtosis
BLUE
Probabilities
n!/(r!)*(n-r)!
Number of ways of giving r objects to n people, such that repetition is allowed
Nr
Question-Counting Principle
In how many ways 3 stocks can be chosen out of 10 stocks in a portfolio? (Combination)
Choosing 3 out of 10 stocks is basically the number of combinations of 3 objects out of 10
Therefore, the number of ways are 10 C 10!
120
3!*(10 3)!
3
In how many ways 3 stocks can be sold, if the sold stock is bought back in the portfolio before the
next stock is sold?
First stock can be sold in 10 ways
Second can be sold in again 10 ways
Third stock can again be sold in 10 ways
Therefore total number of ways become =103 = 1,000
You wish to choose a portfolio of 3 bonds and 4 stocks from a list of 5 bonds and 8 stocks.
How many different 7 asset portfolio can you make from this list.
A. 80
B. 700
C. 1,716
D. 100,800
Solution
B.
5 5! 8! 5* 4 8*7 *6*5
C3 *8 C4 * * 10 * 70 700
3!(5 3)! 4!(8 4)! 2 *1 4 * 3 * 2 *1
Let us consider a Problem: A single 6-sided die is rolled 100 times, and the outcomes are noted..
What is the probability of each outcome? What is the probability of rolling an even number? of
rolling an odd number?
An Experiment is a situation involving chance or probability that leads to results called outcomes.
Example: In the above problem, rolling the dice is the experiment
An outcome is the result of a single trial of an experiment.
Example: when we roll the dice one, the possible outcomes are 1,2,3,4,5 or 6.
An event is a set of outcomes of an experiment (a subset of total possible outcomes).
Example: One event A can be defined as getting an even number, when we roll the dice
A trial constitutes rolling the die once. The experiment can have several trials, like 100 or 200.
Probability: Definitions
A random variable, usually written as X, is a variable whose possible values are numerical
outcomes of a random phenomenon
Examples:
X No of times we receive head when a coin is flipped 5 times
X Average height of a random group of 25 people selected from Bangalore
Discrete Random Variable: A random variable that can take finite number of distinct values
Example: consider an experiment where a coin is tossed 3 times. Random variable X is defined as
X No of heads obtained when we flip a coin 3 times; X {0,1,2,3}
Continuous Random Variable: A random variable that can take an infinite number of possible
values. Usually defined over an interval of values.
Example: consider an experiment where we measure average amount of rainfall received in
Bangalore over a year. Here random variable X is defined as
X any real number between 0 to 200mm
Probability function, P(x), denotes the probability that the discrete random variable X is equal to a
specific value x. P(x) = P(X=x)
W
For any random variable, sum of all the probabilities must equal 1
Example: Random variable X no of heads we receive, when we toss a coin twice
3 Probable outcomes: xi {0,1,2}
P(X=0) = 25%, P(X=1) = 50%, P(X=2) = 25%
Wi) = 1
For continuous random variable, probability of any specific value is zero. Probability is always
calculated for a range of values.
Example: Continuous Random variable X average rain in Bangalore in a year
P(X = 10mm) = 0; (i.e. rain can never be exactly 10mm, it can be 9.999mm or 10.001mm etc)
But P(9.999mm < X < 10.001mm) = 20% (probability is always calculated for a range or interval)
For a Continuous Random Variable, PDF tells us the likelihood of outcomes occurring between
any two pointes. It is used to calculate the probability that outcomes of the continuous
distribution lies within a particular range.
Wl yu) = area under the PDF curve bounded by rl and ru
P(rmin ymax) = 1
Wy
i.e. minimum rain Bangalore may receive is zero, and maximum it can receive is infinity
CDF defines the probability of the continuous random variable being less than a certain value.
Usually denoted as F(x).
F(a) = -a W = area under the PDF curve to the left of point a
&
If x1 < x2, F(x1&2). i.e. CDF is a non decreasing function ( in other words as the upper range of
possible outcomes increases, the cumulative probability also increases or remains same)
F(rmaxWmax ) = 1
P(x > a) = 1 &WW
W& F(a)
Define the probability density function for the price of a zero coupon bond with a notional value of
$10 as
PDF, where x is the price of the bond
Thus P(a ab f(x)dx = (b2 a2)/100
Using the above equation for Probability, calculate
(i) Probability that the price of bond is between 8 and 9? i.e. find P(8
(ii) &W
(iii) W
EduPristine
E FRM-I (2016) 14
Solution
EduPristine
E FRM-I (2016) 15
Some definitions and properties of Probability
Mutually Exclusive events: If one event occurs, then other cannot occur. For a given random
variable, the probability of any of two mutually exclusive events occurring is just the sum of their
individual probabilities. P(A U B) = P(A) + P(B), if A and B are mutually exclusive
Example: Random variable X annual return on portfolio
Event A: return is less then -10%; P(A) = 10%
Event B: return is more than 10%; P(B) = 20%
WhWW
Note: here A and B are exclusive, because, return cannot be more than 10% and less than -10%
simultaneously
Exhaustive: All exhaustive events taken together form the complete sample space
For previous example, if we consider another event C : -10%
Now events A, B and C are mutually exclusive and exhaustive
Thus P(A U B U C) = 1, i.e. P(C) = 70%
Independent Events: One event occurring has no effect on the other event. i.e. we have more
than 1 random variables.
Event A: Kohli scores a century in the next match, P(A) = 10%
Event B: Stock market is up tomorrow, P(B) = 8%
P(A and B) = P(A B) = P(A) x P(B) Also called as Joint Probability
Since the two events are independent, the probability that Kohli gets a century and the Stock
market goes up is 10% x 8% = 0.8%
What if A and B are dependent?
Event A: Kohli scores a century in the next match, P(A) = 10%
Event B: India wins the match, P(B) = 40%
P(A and B) = ?
Conditional probability comes into picture.
Joint Probability
A statistical measure where the likelihood of two events occurring together and at the same point in time are
calculated. Joint probability is the probability of event Y occurring at the same time event X occurs.
Notation for joint probability takes the form:
WyzWyzWhich reads the joint probability of X and Y
The following table shows the joint probability of different events. Lets say an economist is predicting the
market scenario and the price of IBM stock from the next year.
Next year market can be Good, Bad or Neutral
IBM stock may go up or go down
Joint Probability Table:
Market
Good Bad Neutral Total
IBM
UP 10% 30% 5% 45%
DOWN 0% 15% 40% 55%
Total 10% 45% 45% 100%
The probability of IBM stock being Up and Market being Good is 10%
Similarly, the probability of IBM stock being down and Market being neutral is 40%
Conditional Probability
Probability of an event or outcome based on the occurrence of a previous event or outcome. Conditional
probability is calculated by multiplying the probability of the preceding event by the updated probability of
the succeeding event
The probability of event A given that the event B has occurred is P(A/B), which is equal to the ratio of joint
probability of A and B, and unconditional probability of B. P( A B)
P( A / B)
P( B)
The unconditional probability of market being Neutral is 45%. Then using the table below we can
find 3 conditional probabilities.
P(Up/Neutral) = 0.05/0.45
P(Up/Good) = 0.1/0.1
P(Down/Bad) = 0.15/0.45
Joint Probability Table:
Market
Good Bad Neutral Total
IBM
UP 10% 30% 5% 45%
Unconditional Probabilities of
DOWN 0% 15% 40% 55% IBM stock being UP / Down
Total 10% 45% 45% 100%
Market
Unconditional Probability
10% 45% 45%
of Market
Calculate:
Unconditional Probability of market to be good next year?
Conditional Probability of IBM stock rising when the market is neutral?
Conditional Probability of market being good when IBM stock is down?
Question
For a bond with B rating, assume 1 year probability of default for each issuer is 6%, and that
default probability of each issuer are independent. What is the probability that both issuers avoid
default during the 1st year.
A. 88.0%
B. 88.4%
C. 94.0%
D. 96.4%
B.
Both would avoid default only if None defaults
This implies that first does not default AND second does not default
= (1 PD (first)) x (1 PD (second))
= (1 0.06) x (1 0.06) = 0.884 = 88.4%
The probability of happening of event A or event B can be given as the sum of the three portions
defined by the figure below:
P( A) P( A B) P (B )
P( A) P (B )
P( A B) 0
Jensen, a portfolio manager is managing two portfolios. One for High Net Worth Individuals (HNI)
and second for Low Net Worth Individuals (LNI)
HNI portfolio contains 5 bonds and 7 stocks and LNI contains 6 bonds and 11 stocks
One instrument from HNI is transferred to LNI portfolio
What are the probabilities that a stock and a bond is transferred from HNI to LNI?
A. 42%, 58%
B. 58%, 42%
C. 50%, 50%
D. 33%, 67%
Now Jensen selects an instrument from LNI, what is the probability that instrument selected is
stock?
A. 0.5382
B. 0.7821
C. 0.6435
D. None of these
Solution
B.
C.
Here required probability = [P(stock transferred from HNI) AND P(Stock selected from LNI)] OR
[P(bond transferred from HNI) AND P(Stock selected from LNI)]
So, the required probability = (7/12) (11/18) = 139/216 = 0.6435
Hence option C is correct
The unconditional probability of event B is equal to the sum of joint probabilities of event (A,B)
and the probability of event (A,B). Here A is the probability of not happening of A
The joint probability of events A and B is the product of conditional probability of B, given A has occurred and
the unconditional probability of event A
P( B) P ( A B ) P ( Ac B ) P( B / A) P( A) P( B / Ac ) P( Ac )
We know that P(AB) = P(B/A) * P(A)
Also P(BA)= P(A/B) * P(B)
P( B / A) * P( A)
Now equating both P(AB) and P(BA) we get: P( A / B)
P( B)
P(B) can be further broken down using sum rule defined above:
P( B / A) P( A)
P( A / B)
P( B / A) P( A) P( B / Ac ) P( Ac )
Question
Out of a group of 100 patients being treated for chronic back trouble, 25% are chosen at random
to receive a new, experimental treatment as opposed to the more usual muscle relaxant-based
therapy which the remaining patients receive. Preliminary studies suggest that the probability of a
cure with the standard treatment is 0.3, while the probability of a cure from the new treatment
is 0.6.
How many patients (on an average) out of the 100 patients selected at random would be cured?
A. 30
B. 40
C. 37.5
D. 42.5
Some time later, one of the patients returns to thank the staff for her complete recovery.
What is the probability that she was given the new treatment?
A. 0.375
B. 0.425
C. 0.4
D. 0.425
C.
25% are given new treatment =>75% are given old treatment.
WWEWEWKWK
So out of 100 patients 37.5 will get cured
C.
Apply Bayes Theorem
P (New / Cure) = P (Cure / New) * P (New) / P (Cure) = 0.6 * 0.25 / 0.375 = 0.40
Question
Calculate the probability of a subsidiary and parent company both defaulting over the next year.
Assume that the subsidiary will default if the parent defaults, but the parent will not necessarily
default if the subsidiary defaults. Assume that the parent had a 1 year PD = 0.5% and the
subsidiary has 1 year PD of 0.9%.
0.45%
0.5%
0.545%
0.55%
B.
P (S| P) = 1 = P(P & S)/P(P)
P(P & S) = P(P) = 0.5%
Basic Statistics
Basic Statistics
A random variable is characterized by its distribution function. Instead of having to report the
whole function, it is convenient to summarize it by a few parameters, or moments.
The first four moments about the mean:
Mean
Variance
Skewness
Kurtosis
The expected value(Mean) measures the central tendency, or the center of gravity of the
population n
It is given by: x i
P E( X ) i 1
N
A family has 4 members, father, mother and 2 kids Hemal and Rishi who are twins. The average
age of the family members is 20 years. Age of mother and father is 30 and 32 respectively.
Can you tell the age of Hemal?
The graph shows the mean of normal distributions
0.45
0.40
0.35
0.30 Standard Normal Distribution
0.25 P = 0, V = 2
0.20
0.15 P = 1, V = 1
0.10
0.05
0
-4 -2 0 2 4
For a Discrete Random variable, X, with possible values xi, and corresponding probabilities pi,
n
mean is given as
P E( X ) px
i 1
i i
For a Continuous Random Variable X, with possible values xi , and PDF f(x)
x max
P E( X ) xf ( x)dx
x min
Median is the midpoint of the data set when the data is arranged in ascending or
descending order
Half of the observations lie above the median and half are below
Median is important because arithmetic mean can be affected by extremely large or small values
Example Find the median of the following data set:
3, 13, 7, 5, 21, 23, 39, 23, 40, 23, 14, 12, 56, 23, 29
Mode
Mode is the value which occurs most frequently in the data set
A data may have more than one mode or no mode
Example Find the mode for the following data set
11,3,5,11,7,3,11
In the above data set
Number 11 occurs 3 times,
Number 3 occurs 2 times,
Number 5 occurs 1 times,
Number 7 occurs 1 times.
So the number with most occurrences is 11 and is the Mode of this distribution.
Mode = 11
N
PX pX i 1
i i
Expected Value
y Wy yWy
1 0.3 0.03
2 0.1 0.02
3 0.2 0.06
4 0.4 1.6
2.70
E(cX) = E(X) x c
yzyz
yy
E(XY) = E(X) x E(Y) [if X and Y are independent]
(x i P )2
VAR E[( X P ) 2 ] i 1
N
Also, if rearranged
VAR E[( X P ) 2 ] E[ X 2 ] P 2 E[ X 2 ] E[ X ]2
The standard deviation, which is the square root of the Variance, is more convenient to use,
as it has the same units as the original variable X
n
(x i P )2
SD(X) = VAR (x) V i 1
N
Question
Assuming that the distribution of ABC stock returns is a population, what is the population
variance and standard deviation?
05.00
06.80
45.22
80.20
Given two random variables X and Y, what is the Variance of X given Variance[Y] = 100, Variance
[4X - 3Y] = 2,700 and the correlation between X and Y is 0.5?
56.3
113.3
159.9
225.0
B.
The population variance is given by taking the mean of all squared deviations from the mean.
2 = [(12-5.67)2 -5.67)2 -7-5.67)2 -5.67)2 -5.67)2 -5.67)2] / 6 = 45.22 (%2)
The standard deviation is the square root of the variance: 6.72%
D.
Using the theorems on variance and covariance
Variance [4X-zsysz-3)*Var[X]^(1/2)* Var[Y]^(1/2)*correlation[X,Y]
Solve for Var[X] = 225.0
n n n
Var ( X i ) Cov( X , X i j ) Covariance between same variables is also their variance
i 1 i 1 j 1
n n
Var ( X i ) Var ( X ) i
For independent or uncorrelated variables,
i 1 i 1 covariance or correlation = 0
(x
i 1
i P )3
Sk
V3
Symmetric Distribution
Skewness can be negative or positive.
Kurtosis
(x
i 1
i P )4
K
V4
Because of the fourth power, large observations in the tail will have a large weight and hence
create large kurtosis. Such a distribution is called leptokurtic, or fat tailed
0.45 Platykurtic
A kurtosis of 3 is considered average 0.4
K<3
Just as the concept of variance was extended to a set of two variables as covariance, concept of
skewness and cokurtosis can be extended to the set of more variables as coskewness. It is possible
that the set of variables have same mean and variable but not the same skewness and kurtosis
because the ranking of values will be different for these two variables at different point of time
Most risk models ignore the effects of coskewness and cokurtosis
The main reason for this is that as the number of variables increases, the number of coskewness
and cokurtosis terms increase rapidly
BLUE
Unbiased parameter is one for which the expected value of the estimator is equal to the
parameter you are trying to estimate
An unbiased parameter is also efficient if its sampling distribution has minimum variance
An estimator is also consistent if the accuracy of the parameter estimate increases as the sample
size increases
If this estimator is linear, we call it the Best Linear Unbiased Estimator (BLUE)
The random variables X and Y have variances of 2 and 3 respectively, and covariance of 0.5.
dyz
A. 13
B. 29
C. 35
D. 41
Question 2
You are given the following information about the returns of stock P and stock Q:
Variance of return of stock P = 100.0
Variance of return of stock Q = 225.0
Covariance between the return of stock P and the return of stock Q = 53.2
At the end of 1999, you are holding USD 4 million in stock P. You are considering a strategy of
shifting USD 1 million into stock Q and keeping USD 3 million in stock P. What percentage of risk,
as measured by standard deviation of return, can be reduced by this strategy?
A. 0.50%
B. 5.00%
C. 7.40%
D. 9.70%
Which type of distribution produces the lowest probability for a variable to exceed a specified
extreme value X which is greater than the mean assuming the distributions all have the same
mean and variance?
A. A leptokurtic distribution with a kurtosis of 4
B. A leptokurtic distribution with a kurtosis of 8
C. A normal distribution
D. A platykurtic distribution
Question 4
1. B.
syzsysz
Var(X - zsysz-2*Cov(x,y)
Var(cX) = c^2 * Var(X)
Cov (ax,by) = abCov(x,y)
^syz2 sy2 sz
2. B.
5.00%
Solution
3. D.
By definition, a platykurtic distribution has thinner tails than both the normal distribution and any
leptokurtic distribution. Therefore, for an extreme value X, the lowest probability of exceeding it will be
found in the distribution with the thinner tails
A. Incorrect. A leptokurtic distribution has fatter tails than the normal distribution. The kurtosis indicates
the level of fatness in the tails, the higher the kurtosis, the fatter the tails. Therefore, the probability of
exceeding a specified extreme value will be higher
B. Incorrect. Since answer A. has a lower kurtosis, a distribution with a kurtosis of 8 will necessarily produce
a larger probability in the tails
C. Incorrect. By definition, a normal distribution has thinner tails than a leptokurtic distribution and larger
tails than a platykurtic distribution
4. D.
In a skewed distribution, the mean is pulled in the direction of the extreme scores or tail (same as the
direction of the skew), and the median is between the mean and the mode.
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
Quantitative Analysis - II
EduPristine www.edupristine.com
Agenda
Probability distributions
Discrete vs. Continuous probability distribution
Discrete probability distribution: Bernoulli, Binomial, Poisson
Continuous Probability Distribution
Sampling
Central Limit Theory
Hypothesis testing - Part 1
Probability Distributions
Binomial Distribution
Assumptions:
There are n trials
Each trial has two possible outcomes, success or failure
The probability of success p is the same for each trial
Each trial is independent
If we take n Bernoulli trials, and say out of those n trials we have total number of x successes,
then the probability of such an event can be given as:
P( X x) C x * p x * (1 p ) ( n x )
n
1. There are 10 bonds in a credit default swap basket; the probability of default for each of the
bonds is 5%. The probability of any one of the bond defaulting is completely independent of what
happens to the other bonds in the basket. What is the probability exactly one bond default?
A. 5%
B. 50%
C. 32%
D. 3%
2. Company ABC was incorporated on January 1, 2004. it has expected annual default rate of 10%.
Assuming a constant quarterly default rate, what is the probability that company ABC will not
have defaulted by April 1, 2004?
A. 2.4%
B. 2.5%
C. 97.4%
D. 97.5%
3. A corporate bond will mature in 3 years. The marginal probability of default in year one is 0.03%.
The marginal probability of default in year 2 is 0.04%. The marginal probability of default in year 3
is 0.06%. What is the cumulative probability that default will occur during the 3 year period?
A. 0.1247%
B. 0.1276%
C. 0.1299%
D. 0.1355%
Solution
1. C.
One particular bond defaults and other nine do not with the probability 0.05* (0.95)^9 can happen in 10
different ways
= 10 * 0.05^1* (0.95)^9 = 32%
2. C.
Let the probability of not defaulting in 1 quarter is (nd). Then the probability of not defaulting for a full year is
(nd)4. This implies that the probability of defaulting within 1 year time is {1-(nd)4}, which is given as 10%.
1-(nd)4 = 0.1 which implies
(nd) = 0.91/4
= 97.4%
3. C.
The cumulative probability of default= 1-{Product of marginal probabilities of not defaulting
= 1-{(1-0.0003)*(1-.0004)*(1-0.0006)}
= 0.001299
Therefore the cumulative probability of default is 0.1299%
Ox O
e xt0
P( X x) x!
0 otherwise
0.40
=1 0.20
=5
0.35 0.18
0.16
0.30
0.14
0.25 0.12
0.20 0.10
0.15 0.08
0.06
0.10
0.04
0.05
0.02
0.00 0.00
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
2. The number of false fire alarms in a suburb of Houston averages 2.1 per day. What is the
(approximate) probability that there would be 4 false alarms on 1 day?
A. 1.0
B. 0.1
C. 0.5
D. 0.0
Solution
1. C.
The Normal distribution can approximate the distribution of a Poisson random variable with a large lambda
d
INCORRECT: A, The mean of a Poisson distribution must be large to allow approximation with a Normal
distribution
INCORRECT: B, The variance of a Poisson distribution must be large to allow approximation with a Normal
distribution
INCORRECT: D, The Normal distribution can approximate the distribution of a Poisson random variable with a
2. B.
Use Poisson distribution
P(X = x) = [O x *e- O]/ O!
There is a city which hosts two taxi-cab companies, the Blue Cab Co. and the Green Cab Co. Blue
cabs are blue and Green cabs are green; they are otherwise identical. 70 percent of the cabs in the
city are Blue cabs, and 30 percent of the cabs in the city are Green cabs. Moreover, historically
speaking, Blue cabs have been involved in 70% of all traffic accidents in the city that involved cabs,
and Green cabs have been involved in 30% of all traffic accidents in the city that involved cabs.
One night, there is a traffic accident involving a taxi-cab in the city, to which there is one witness.
Authorities perform extensive tests on the witness, and determine that his ability to recognize
cabs by their color at night is approximately 80 percent accurate and 20 percent inaccurate
(meaning that when he is wrong he does not say he doesn't know, but rather misidentifies it as
being of the other color). The witness says the taxi-cab involved in the accident was 'blue.' On
these facts, and strictly assuming the taxi-cab was not from some other city, what is the
approximate probability that the taxi-cab involved in the accident belonged to the Blue Cab Co.
0.7 0.3
Witness
Identifies it as: =0.06
Therefore the probability of the car being actually blue, when the witness identified it as blue
equals: (0.56/0.62) 0.903
Solution
Initially Jack has 3 white balls and 2 red balls in his box while his friend Andrew has 4 white and 5
red balls
After the exchange Jack has 2 white and 3 red balls, and Andrew has 5 white and 4 red balls.
Therefore the probability of picking a red ball from Andrews box is:
P(RAndrew
Now Tom stole a white ball from one of the two boxes. To make a calculated guess about who lost
1 white ball, we need to calculate the conditional probabilities
P(Jacks box/If the balls is White)= Probability of white balls in Jacks box/(Probability of white ball
:W
2/5
18 / 43
2 / 5 5 / 9 5/9
Similarly, P(Andrews box / White) 25 / 43
2 / 5 5 / 9
Point to note here is that the white ball can come from 2 boxes only, so the sum of conditional
our case
Bonds
Upgrade No change Downgrade Total
Stocks
Outperfom 10% 30% Y C
Underperform X 15% 35% 55%
Total A 45% B 100%
Solution
A. X = 5%, Y=5%, A=15%, B=40%, C=45%
C. p(stocks outperform)
= P(stocks outperform/Bond downgraded) *P(Bond downgraded)
WW
WW
Uniform distribution
The simplest continuous distribution function is the uniform distribution. This is defined over a
The density function is:
Its mean and variance are given by:
y
sy2 / 12
1
b-a
a b
Assume we use continuous uniform distribution U(0,10) to generate a series of random numbers.
Which of the following statements is Correct?
A. The number 5 is likely to be observed much more often than any other number
B. Numbers between 4 and 6 are more likely to occur than the number between 6 and 10, because the first
interval is closer to the center of the distribution
C. Numbers between 1 and 3 are as likely as the number between 4 and 6
D. Numbers between 1 and 3 are less likely than the number between 4 and 6, due to skewness of
the distribution
Solution
C.
The normal distribution is defined by first two moments, mean () and variance (2)
The probability density function P(x) of normally distributed variable is given by:
1 ( x P )2
P( x) exp
2SV 2 2V 2
b
P ( a d X d b)
The probability of the value lying between a and b is given by:
P( x).dx
a
The probabilities of outcomes further above and below the mean get smaller and smaller, but do
not go to zero (extent to infinity)
It is a CDF, with total area under the curve equal to 1.
Confidence interval, is the range of values around the expected outcome, within which we expect
the actual outcome to lie in, during specified percentage of time.
Confidence Level (1 ) Confidence interval
68% -1 to 1
90% -1.65 to 1.65
95% -1.96 to 1.96
99% -2.58 to 2.58
Summation property: The property states that, assuming there is a collection of normal
distributions:
The sum of the means of all the independent normal distributions form a normal distribution
The sum of the variances of all the independent normal distributions form a normal distribution
68% of data
95% of data
99.7% of data
The standard normal distribution has mean = 0 and standard deviation sigma = 1
Consider a stock with an initial price of $100. Its price one year from now is given by
S = 100exp(r), where the rate of return r is normally distributed with a mean of 0.1 and a
standard deviation of 0.2. With 95% confidence, after rounding, S will be between
$67.57 and $147.99
$70.80 and $149.20
$74.68 and $163.56
$102.18 and $119.53
C.
Note that this is a two-
t
dsd
s 0.2) = $100 * exp(0.492) = $163.56
Question
Let X be a uniformly distributed random variable between minus one and one so that the standard
deviation of X is 0.577. What percentage of the distributions will be less than 1.96 standard
deviations above the mean:
A. 100%
B. 97.5%
C. 95%
D. Insufficient information provided
A
The answer requires understanding of distributions and standard deviation. The key is that every distribution
has a standard deviation. However the number of standard deviations associated with different probabilities
are different for each distribution. In this case 1.96 standard deviation represents a move of 1.12 or less. As
the total distribution is defined as falling between minus one and one the correct answer is A
Question
D.
In the diagram given below, the area representing the region P(-W
d-
P(-WWW
Hence option D is correct
Lognormal distribution
A random variable X is said to have a lognormal distribution if its logarithm Y = ln(X) is normally
distributed
1 (ln( x) P ) 2
The lognormal density function has the following expression: P(x) exp
x 2SV 2 2V 2
(100,2) (2,0.3)
1.2
0.6
(a) 0.9 (b)
0.4
Density
0.6
0.2
0.3
0.0 0.0
0 100 200 300 400 1.0 1.5 2.0 2.5 3.0
Original Scale Log Scale +
If there are k independent normal variables, than the sum of their squares has a chi squared
distribution.
The chi-square distribution is a family of distributions, depending on degrees of freedom: d.f. = n
1. Mean = k, and variance = 2k
Assymmetrical, but approaches normal distribution as degrees of freedom increases. As the
variable is sum of squared values, it can only have non-negative values.
0 4 8 12 16 20 24 28 F2 0 4 8 12 16 20 24 28 F2 0 4 8 12 16 20 24 28 F2
Assume you have empirical data showing historical returns (v) for a given financial variable
(e.g.: Forex rate), how could you perform a quick test of the validity of the power law
Prob(v > x) = K * x-a where x is large, as a good model of the tail of the distribution?
A. Plot the probability of v exceeding x standard deviations against x
B. Plot the probability of v exceeding x standard deviations against Log of x
C. Plot the Log of the probability of v exceeding x standard deviations against x
D. Plot the Log of the probability of v exceeding x standard deviations against the Log of x
D.
The mathematical relationship in the question can be rewritten (by taking the logs on both sides): Log(Prob(v
> x)) = Log(K) aLog(x), i.e. the plot of the Log of the probability of v exceeding x standard deviations against
the log of x should be a straight (decreasing) line if the relationship strictly holds.
The intercept is an estimate of Log of K and the slope of the line yields the parameter a.
A probability sample is a sample selected such that each item or person in the population being
studied has a known likelihood of being included in the sample
The sampling distribution of the sample mean is a probability distribution consisting of all
possible sample means of a given sample size selected from a population
Need for Sampling:
The physical impossibility of checking all items in the population
The cost of studying all the items in a population
The sample results are usually adequate
Contacting the whole population would often be time-consuming
The destructive nature of certain tests
Suppose theres a population of 4 oldest scientists in a university: Jack, Andrew, Michelle and Tom
Random variable, X is the ages of the individuals 79
Ages of Population
78
Values of X: 78, 76, 72, 74 77
76
75
74
Summary Measure for Population Distribution 73
72
71
N 70
X
69
i Andrew Jack Michelle Tom
i 1
Average Age
N Prob. of selection
0.3
78 76 72 74
75 0.25
4 0.2
N 0.15
X
2
0.1
i
i 1
2.236 0.05
N 0
Andrew Jack Michelle Tom
0.30
0.25
0.20
0.15
0.10
0.05
0
72 73 74 75 76 77 78
X Px
N
2
Vx i 1
i
72 752 73 752 78 752 1.58
N 16
V2
Variance of sampling distribution of mean (sqr of standard error) s2
n
Suppose the standard deviation of a normal population is known to be 10 and the mean is
hypothesized to be 8. Suppose a sample size of 100 is considered. What is the range of sample
means that allows the hypothesis to be accepted at a level of significance of 0.05?
A. Between -11.60 and 27.60
B. Between 6.04 and 9.96
C. Between 6.355 and 9.645
D. Between -8.45 and 24.45
Solution
B.
d-1.96 and 1.96
y- 8) / (10 / Sqrt (100)) and -1.96 <=z<=1.96
which implies that the sample mean X must be between 6.04 and 9.96
Review of Statistics
Statistical Inference
Hypothesis Testing
Type I Errors, Type II Errors, p-value (1 hr)
One-Tailed Test, Two-Tailed Test (2 hrs)
Statistical inference
Statistical Inference is the process of drawing a relationship between a population and a sample
drawn from that population
Estimation & Hypothesis Testing: (Branches of statistical inference)
Estimation: Finding out estimator values (mean, variance etc) of the sample
Hypothesis testing: Judge whether the hypothesis made is reliable or not based on the sample estimators
A statistical hypothesis test is a method of making statistical decisions from and about
experimental data
Null-hypothesis testing answers the question:
How well the findings fit the possibility that chance factors alone might be responsible"
Example: Does your score of 6/10 imply that I am a good teacher???
Linearity
Unbiasedness
Minimum Variance
Efficiency
Best Linear unbiased estimator (BLUE)
Consistency
Unbiased estimator: One or more values of an estimator is equal to the true value of a parameter
Efficient estimator: Considering only the unbiased estimators of a parameter, the one which has
least variance is called the efficient estimator
Consistent estimator: The estimator which approaches the true value of its parameter as the
sample size increases
Christos, brand manager for a leading financial training center, wants to introduce a new niche finance
course for MBA students. He met some industry stalwarts and found that with the skills acquired by
attending such a course, the students would be very hot in job market
He meets a random sample of 100 students and discovers the following characteristics of the market
Mean household income to $20,000
Interest level in students = high
Current knowledge of students for the niche concepts = low
Christos strongly believes the course would adequately profitable in students if they have the buying
power for the course. They would be able to afford the course only if the mean household income is
greater than $19,000
Would you advice Christos to introduce the course?
What should be the hypothesis?
Hint: What is the point at which the decision changes (19,000 or 20,000)?
What about the alternate hypothesis?
What other information do you need to ensure that the right decision is arrived at?
Hint: confidence intervals / significance levels?
Hint: Is there any other factor apart from mean, which is important? How do I move from population parameters to
standard errors?
What is the risk still remaining, when you take this decision?
Hint: Type-I / II errors?
Hint: P-value
To reach a final decision, Christos has to make a general inference (about the population) from
the sample data
Criterion: Mean income across all households in the market area under consideration
If the mean population household income is greater than $19,000, then Christos should introduce the
product line into the new market
Christoss decision making is equivalent to either accepting or rejecting the hypothesis:
The population mean household income in the new market area is greater than $19,000
The term one-tailed signifies that all z-values that would cause Christos to reject H0, are in just
one tail of the sampling distribution
-> Population Mean
H0: d $19,000
Ha: ! $19,000
0.25
0.2
Critical Value (Xc)
0.15
0.1
0.05
$19,000
Sample mean values greater than $19,000--that is x-values on the right-hand side of the sampling
distribution centered on = $19,000--suggest that H0 may be false
More important the farther to the right x is , the stronger is the evidence against H0
Reject H0 yc
Standard deviation for the sample of 100 households is $4,000. The standard error of the mean
(sx) is given by:
s
sx $400
n
Critical mean household income xc through the following two steps:
Determine the critical z-value, zc. For D = 0.05:
zc = 1.645
Substitute the values of zc, s, and P (under the assumption that H0 is "just" true )
xc = P + zcs = $19,658
Decision Rule:
If the sample mean household income is greater than $19,658, reject the null hypothesis and
introduce the new course
The value of the test statistic is simply the z-value corresponding to = $20,000
xP 0.25
Z 2.5
sx
0.2
Xc P Z c *V 0.15
0.1
Errors in estimation
Significance Level
D -> Significance level
the upper-bound probability of a Type I error
1 - D ->confidence level
the complement of significance level
P-value
The probability of obtaining an observed value of 0.25
x (from the sample) is as high as $20,000 or more
when actual populations mean (P) is only 0.2
$19,000 = 0.00621
0.15
This value is sometimes called the actual
significance level, or the p-value
0.1
Calculated probability of rejecting the null
hypothesis (H0) when that hypothesis (H0) is true
0.05
(Type I error)
The actual significance level of 0.00621 0
in this case means that the odds are less than p-value = 0.00621
= $19,000
62 out of 10,000 that the sample mean
income of $20,000 would have occurred
entirely due to chance Do not Reject H0 Reject H0
(when the population mean income
is $19,000)
Decision Rule: smallest level of significance for which the null hypothesis can be rejected
Question
A. I only
B. I and IV
C. I and III
D. I, II and IV
A.
II is false. A decrease in the level of Type-I error increases the Type-II error
III is false. Type I error is when a true null hypothesis is rejected
Question
Consider an exam taken by 15,000 FRM candidates. Mean score for the exam was 64 for all the
6,400 candidates who studied at least 250 hours in the preparation of the exam. Assuming a
population standard deviation of 16. What would be 99% confidence interval for the mean score
'0.005 =2.575)
I. 64 0.52
II. 64 0.80
III. 64 5.15
IV. 64 8.05
A.
d
= 64 2.575 * 16/sqrt(6400)
= 64 0.515
What if Christos surveyed the market and found that the student behavior is estimated to be:
They would found the training too expensive if their household income is < US$19,000 and hence would
not have the buying power for the course?
They would perceive the training to be of inferior quality, if their household income is > US$19,000 and
hence not buy the training?
How would the decision criteria change? What should be the testing strategy?
Hint:
From the question wording infer: Two tailed testing
Appropriately modify the significance value and other parameters
h-test
Appropriate change in the decision making and testing process:
Students will not attend the course if:
The household income >$19,000 and the students perceive the course to be inferior
The household income is <$19,000
This becomes a two tailed test wherein the student will join the course only when the household lie
between a particular boundary. i.e. the household income should be neither very high neither very low
What if the sample had not been large enough?? For example, if Christos had met only 25
students, then what?
Conduct t-Test when sample size is small
Let the sample size, n = 25, X = $20,000, s = $8,000
From the t-table tc = 1.71 for D = 0.05 and d.f. = 24
Decision rule: Reject H0 if t ! 1.7l.
Points to observe:
You could not launch the course.. Why?
Hint: Is it because of T-Test?
NO!
Its because the sample size is small =>
Less value of n =>
Higher standard error =>
Lower confidence in rejecting the hypothesis =>
Almost akin to not taking a decision (hence not launching the product)
T- Test to be conducted, when sample size (n) is small (Typically<30) Degrees of freedom = (n-1)
Christos has surveyed the market and decided to launch the course. He has two markets in mind,
where he can launch the course (and hence conducts the survey):
Chicago
Mean Income (Sample Size = 100): $19,500
Standard Deviation(s1): $300
New York:
Mean Income (Sample Size = 100): $18,500
Standard Deviation(s2): $400
What if Christos wants to launch the course in one of the markets?
What would be the decision criteria? What should be the testing strategy?
hc n
tc n = 0
The only treatment to be made different is that the standard error has to be calculated as:
/c n h
Then the test can be appropriately modified 0.25
as a two-h,
0.2
h -> Population Mean
H0h = $0 = 0.025
0.15
= 0.025
Hah
Since we are checking for significance 0.1
H0h = $1,000
Hah
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
EduPristine www.edupristine.com
Agenda
In testing for variances, there are two different tests, because sum of two chi-squares is not a
chi-square
H0: 2 02
HA: 2 > 02
D
(n 1)s 2
F 2
Where
F2 = standardized chi-square variable
n = sample size
s2 = sample variance
2 = hypothesized variance
F2
D D/2
D/2
F2 F2
Question
A brand new air conditioner from Carrier must maintain the selected temperature with little
variation. Specifications from the company call for a standard deviation of not more than 2
degrees (or variance of 4 degrees2). A sample of 4 air conditioners is tested and yields a sample
variance of s2 = 6. Test to see whether the standard deviation specification is exceeded.
Use D = 0.05
D = 0.05
H0: 12 22
HA: 12 22 > 0
D
s12
F
s22
Although its not a rule but always take the higher sample variance as the numerator and lower
sample variance as the denominator
s 22 = Variance of Sample 2
(n2 1) = denominator degrees of freedom
EduPristine
E FRM-I (2016) 138
The F distribution
In the F table:
Numerator degrees of freedom determine the row
Denominator degrees of freedom determine the column
s12
F where df1 = n1 1 ; df2 = n2 1
s 22
H 0: 12 22 t 0 H 0: 12 22 = 0
H A : 12 22 < 0 HA: 12 22
H0: 12 22
H A : 12 22 > 0
D D/2
0 F 0 F
F-test: Question
You are a financial analyst for a brokerage firm. You want to compare dividend yields between
stocks listed on the Dow30 & EURO STOXX 50. You collect the following data:
Dow30 hZK^dKyy
Number 30 50
Mean 3.27 2.53
Std dev 1.5 1.4
Is there a difference in the variances between the Dow30 & EURO STOXX 50 at the D = 0.05 level?
0 F
F = 1.148 is not greater than the critical F value of 1.881, so we do not reject H0
Chebyshev's inequality says that at least 1 - 1/k2 of the distribution's values are within k standard
deviations of the mean.
Where k is any positive real number greater than 1
Monitoring Correlation
EduPristine
d i i FRM-I (2016)
(20 6) 147
Linear regression with one Regressor
Scatter plot
Visual representation of relationship between dependent and independent variable
Dependent/Independent variables
The main goal of regression analysis is to measure how the changes in one variable known as dependent
variable can be explained by the changes in one or more other variables called the independent variables
Linear relationship
There is linear relationship between dependent and independent variables
Parameters
Slope and intercept term of the regression equation are know as parameters
Error term
Difference between expected value from the equation and actual value. This is mainly because the
dependence of the dependent variable on other factors is not taken into account
Middle
Hours Mumbai Delhi Chennai Kolkata Bangalore Pune Hyderabad Online Singapore
East
10 20 7 5 13 10 11 14 9 7 12
20 8 24 34 24 16 19 20 20 25 12
30 19 8 16 37 62 29 33 25 36 30
40 67 31 44 43 32 19 38 27 49 35
50 36 46 78 57 36 82 55 33 53 41
60 67 54 90 58 23 45 62 67 58 78
70 56 68 93 71 76 72 68 81 70 57
80 81 89 7 6 45 68 83 58 90 9
The population
120
100
80
Marks in Test
60
40
20
0
0 10 20 30 40 50 60 70 80 90
Hours of Study
z
Y E 0 E1 X u
Observed Value of
zyi
Slope = 1
Wsz Random Error for this
yi x value
ui
Intercept = 0
xi x
Y 0 1 X u
Linear component Random Error
component
Hours Mumbai
10 20
20 8
30 19
40 67
50 36
60 67
70 56
80 81
z
y b 0 b1 x e
Observed Value of
zyi
ei
Slope = 1
Wsz Random Error for
yi this x value
Intercept = 0
xi x
yi b 0 b1x e
Represents the influence of all the variable which we have not accounted for in the equation
It represents the difference between the actual y values as compared the predicted y values
from the Sample Regression Line
Wouldnt it be good if we were able to reduce this error term?
What are we trying to achieve by Sample Regression?
Our objective
Y 0 1 X u
y i b 0 b1x
e 2
(y y) 2
(y (b 0 b1 x)) 2
The sum of the residuals from the least squares regression line is 0
( y y ) 0
The simple regression line always passes through the mean of the y variable and the mean of the
x variable
d0 1
b1
( x x )( y y )
(x x) 2
xy n
x y b0 y b1 x
b1
x
2
( x)
2
b0 is the estimated average value of y when the value of x is zero. More often than not it does not
have a physical interpretation
b1 is the estimated change in the average value of y as a result of a one-unit change in x
y
Y b0 b1 X
slope of the line(b1)
b0
x
The underlying relationship between the X variable and the Y variable is linear
For a given value of Xi the sum of error terms is equal to 0
The error term is uncorrelated with the explanatory variable X
Error values are normally distributed for any given value of X
The probability distribution of the errors for a given Xi is normal
The probability distribution of the errors for different Xi has constant variance (homoscedacity)
Error values u for given Xi are statistically independent, their covariance is zero
X
Once we fulfill these assumptions in Linear Regression , we are able to estimate the variance and
standard errors of b0 and b1 and this has been possible because of the properties of OLS method
y
yi RSS = Residual sum of squares
y
TSS = Total sum RSS = (yi - yi )2
of squares _
TSS = (yi - y)2
y _
_ ESS = (yi - y)2
_
y y
ESS = Explained Sum of squares
x
yi
TSS ( y y) 2
RSS ( y y ) 2
ESS ( y y ) 2
Where:
y = Average value of the dependent variable
y = Observed values of the dependent variable
y = Estimated value of y for the given x value
The coefficient of determination is the portion of the total variation in the dependent variable
that is explained by variation in the independent variable
The coefficient of determination is also called R-squared and is denoted as R2
2 SSR
R
SST
where
0 d R2 d1
Coefficient of determination
SSR sum of squares explained by regression
R2
SST total sum of squares
R2 r2
Where:
R2 = Coefficient of determination
r = Simple correlation coefficient
R2 = 1
Perfect linear relationship between x and y:
R2 = 1 x
100% of the variation in y is explained by variation in x
R2 = +1 x
0 < R2 < 1
Weaker linear relationship between x and y:
x
Some but not all of the variation in y is explained by
variation in x
y
y
R2 = 0
No linear relationship between x and y:
dz
(None of the variation in y is explained by variation in x)
R2 = 0 x
Question
Paul Graham, FRM is analyzing the sales growth of a baby product launched three years ago by a
regional company. He assesses that three factors contribute heavily towards the growth and
comes up with the following results:
zy1 y2 y3
Sum of Squared Regression [SSR] = 869.76
Sum of Squared Errors [SSE] = 22.12
Determine what proportion of sales growth is explained by the regression results.
A. 0.36
B. 0.975
C. 0.64
D. 0.55
What should be the answer if Instead of Sum of Square Errors, we had Standard Error of Estimate
(SEE)????
The result of the linear regression is: Y = 0.10 - 0.50 X with a correlation coefficient R = (-0.90).
The fraction of the variance of Y attributable to X is equal to:
A. (-0.90)
B.
C.
D. (-0.50)
Solution
C.
R-squared is the square of the correlation coefficient and measures the fraction of the variance of Y that is
attributable to X.R2 = (-0.90)2 = 0.81
y y y
r = -1 x r = -0.6 x r=0 x
y y
r = +.3 x r = +1 x
r
( x x )( y y )
[ ( x x ) ][ ( y y )
2 2
]
n xy x y
r
[n( x 2 ) ( x )2 ][n( y 2 ) ( y )2 ]
Where:
r = Sample correlation coefficient
n = Sample size
x = Value of the independent variable
y = Value of the dependent variable
How do we know whether the values of b0 and b1 that we have found are actually meaningful?
It is actually possible that our sample was a random sample and it has given us a totally wrong
regression line?
We do know a lot about the sample error term e but what do we know about the error terms
u of the Population Regression Function?
How do we proceed from here?
The standard deviation of the variation of observations around the regression line is estimated
by:
RSS
su
n k 1
Where:
RSS = Residual Sum of Squares (summation of e2)
n = Sample size
k = number of independent variables in the model
Xi su
2
s bo
n (x x) 2
su su
s b1
(x x) 2
( x) 2
x 2
n
Where:
sb1 = Estimate of the standard error of the least squares slope
RSS
su n2 = Sample standard error of the estimate
Variation of observed y values from the Variation in the slope of regression lines from
regression line different possible samples
y y
y y
d.f. n2
Where:
b1 = Sample regression slope coefficient
,
sb1 = Estimator of the standard error of the slope
Question
A risk analyst performs a simple linear regression on return data comprising three variables
evolving in time and obtains, amongst others, the following statistics:
Based on these data at a 95% confidence level, the analyst should conclude that:
A. The intercept and X Variable 2 are statistically significant
B. X Variable 1 and X Variable 3 are statistically significant
C. X Variable 1, X Variable 2 and X Variable 3 are all statistically not significant
D. More information is required, such as the corresponding p-values, before any meaningful deductions may be
made
A.
A is correct. (Relatively) small standard errors and high t-stats are one indication of indicate statistical
significance
B is incorrect. (Relatively) large standard errors and low t-stats are one indication of indicate statistical
significance
C is incorrect. Negative t-stats are not an indication of statistical insignificance
D is incorrect. The p-values are redundant information if the t-stat is provided. That is, the p-values tell one
nothing more than the t-stats do
One of the important assumptions of Linear regression is that the variance of the error term is
same across all observations. This type of regression is called homoscedastic regression
When the requirement of a constant variance is violated, we have a condition of
heteroscedasticity
Error u
Predicted y
Effects of Heteroscedasticity
Regression coefficients are not affected
Standard errors are usually unreliable
Hypothesis testing of regression coefficient becomes insignificant
This theorem says that if the assumptions of linear regression are true, then the OLS estimators
have following properties:
OLS estimated coefficients have minimum variance compared to other methods of estimation
OLS estimated coefficients are based on linear functions
OLS estimated coefficients are unbiased
OLS estimate of variance of errors is unbiased
Multiple Regression
There exists a linear relationship between the dependent and independent variables
The expected value of the error term, conditional on the independent variables is zero
The error terms are homoskedastic, i.e. the variance of the error terms is constant for all the
observations
The expected value of the product of error terms is always zero, which implies that the error
terms are uncorrelated with each other
The error term is normally distributed
The independent variables doesn't have any linear relationships between each other
In simple linear regression, the dependent variable was assumed to be dependent on only one
variable (independent variable)
In General Multiple Linear Regression model, the dependent variable derive sits value from two or
more than two variable
General Multiple Linear Regression model take the following form:
Yi b0 b1 X 1i b2 X 2i ......... bk X ki H i
Where:
Yi = ith observation of dependent variable Y
Xki = ith observation of kth independent variable X
b0 = intercept term
bk = slope coefficient of kth independent variable
i = error term of ith observation
n = number of observations
k = total number of independent variables
As we calculated the intercept and the slope coefficient in case of simple linear regression by
minimizing the sum of squared errors, similarly we estimate the intercept and slope coefficient in
multiple linear regression
n
Sum of Squared Errors H
i 1
i
2
is minimized and the slope coefficient is estimated.
Intercept Term (b0): It's the value of dependent variable when the value of all independent
variables become zero
b0 Value of Y
when X 1 X2 ....... X k 0
Slope coefficient (bk): It's the change in the dependent variable from a unit change in the
corresponding independent (Xk) variable keeping all other independent variables constant
In reality when the value of the independent variable changes by one unit, the change in the dependent
variable is not equal to the slope coefficient but depends on the correlation among the independent
variables as well
Therefore, the slope coefficient are called partial slope coefficients as well
Coefficient of determination(R2) can also be used to test the significance of the coefficients
collectively apart from using F-test
SST - SSE RSS Sum of Squares explained by regression
R2
SST SST Total Sum of Squares
The drawback of using Coefficient of determination is that the value of the coefficient of
determination always increases as the number of independent variables are increased even if the
marginal contribution of the incoming variable is statistically insignificant
To take care of the above drawback, coefficient of determination is adjusted for the number of
independent variables taken. This adjusted measure of coefficient of determination is called
adjusted R2
Adjusted R2 is given by the following formula:
Where: Ra2
n 1
1 2
u 1 R
n k 1
n = Number of Observations
k = Number of Independent Variables
Ra2= Adjusted R2
EduPristine
E FRM-I (2016) 198
Multicollinearity
Multicollinearity refers to the condition when two or more independent variables in a multiple
regression are highly correlated with each other
There are two types of multicollinearity:
Perfect multicollinearity
One independent variable has perfect correlation with other independent variables
In this case, coefficients of the model cannot be determined
Imperfect multicollinearity
One or more independent variable is correlated to some degree (not perfect) to other independent variables
In this case, coefficients can be determined but might result in errors during test of statistical significance
(chances of Type II error)
Effects of Multicollinearity
There is greater probability that we will incorrectly conclude that a variable is not statistically significant (A type II
error)
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
Quantitative Analysis-IV
EduPristine www.edupristine.com
Agenda
The values of the slope coefficients doesn't tell anything about their significance in explaining the
dependent variable
Even an unrelated variable when regressed would give some value of slope coefficients
To exclude the cases where the independent variables doesn't significantly explain the dependent
variable, we need the hypothesis testing of the coefficients for checking whether they contribute
in explaining the dependent variable significantly or not
The t-statistic is used to check the significance of the coefficients
The t-statistic used for the hypothesis testing is same as used in the hypothesis testing of
coefficient of simple linear regression
Following are the hypothesis and alternative hypothesis to check the statistical significance of bk:
Hypothesis H0: bk =0
Alternative Hypothesis (Ha): bk
The t-statistic
of (n-k-1) degrees of freedom for the hypothesis testing of the coefficient bk
bk bk
t
s
bj
If the value of t-statistic lies within the confidence interval, H0 can't be rejected
Where:
tc is the critical t-value, and
sb is the standard error
j
The regression equation can be used for making predictions about the dependent variable by
using forecasted values of the independent variables
Yi b0 b1 X 1i b2 X 2i ......... bk X ki
Where:
F-Statistic
An F-test explains how well the dependent variable is explained by the independent variables
collectively
In case of multiple independent variable, F-test tells us whether a single variable explains a
significant part of the variation in dependent variable or all the independent variables explain the
variability collectively
Suppose Edward now wants to check whether the household income and expenses collectively
explains the variation in his pocket money
The following hypothesis can be formed:
H0: HI=HE=0 versus Ha,/,
RSS
MSR k
F-statistic is given as: F
MSE SSE
Where: n k 1
EduPristine
E FRM-I (2016) 207
F-Statistic (cont.)
Decision rule for F-test: Reject H0 if the F-statistic > Fc (Critical Value)
The numerator of F-statistic has degrees of freedom of "k" and the denominator has the degrees
of freedom of "n-k-1"
If H0 is rejected then at least one out of two independent variable is significantly different that
zero.
This implies that at least one out of household income(independent variable) or household
expenses(independent variable) explains the variation in the pocket money of Edward
Restricted regression is one in which coefficients of some independent variable are assumed to be
zero
Therefore, any regression with an omitted variable is restricted regression (as coefficients of an
omitted variable are assumed to be zero)
Model Misspecification
3.5
3 [SERIES
Penalty Factors
2.5 NAME],
When the true model is not among the models considered, the probability of selecting the best approximation
to the true DGP approaches 1 as the sample size gets large
AIC is although inconsistent but is asymptotically efficient, whereas the SIC is NOT
Characterizing Cycles
Covariance Stationary A series which have its mean and its variance structure stable over time
Covariance Stationary
Covariance stationary places no restrictions on other aspects of the distribution of the series
Autocorrelation Function
Application (Example)
When building forecasting models, dont pretend that the model is true. Instead, be aware that the
model is approximating a more complex reality.
These models vary in their specifics and have different strengths in capturing different sorts of
autocorrelation behaviour.
The key feature here is the sharp cut off in the autocorrelation function. All the autocorrelation are
0 beyond displacement 1, the order of the MA process.
When the MA process is inverted, it can express the current value of the series not in terms of a
current shock and lagged shock but rather in terms of a current shock and lagged values of the
series
The partial autocorrelation function for the AR(1) process cuts off abruptly
Partial autocorrelation are just the last coefficients in a sequence of successively longer population
autoregressive. The 1st partial correlation is just the autoregressive coefficient, and coefficients on
all longer lags are 0
Estimation of Volatility
Let xi be the continuously compounded return during day i (between the end of day
i-1 and end of day I)
>n be the volatility of the return on day n as estimated at the end of day n-1
Variance estimate for next day is usually calculated as:
variance = average squared deviation from average return over last n days
x
n
2
i x
i 1
Variance
n 1
Mean of returns (x-bar) is usually zero, especially if returns are over short-time period
(say, daily returns). In that case, variance estimate for next day is nothing but simple average
(equally weighted average)
n of previous n days squared returns
x
2
i
i 1
Variance
n 1
What if the volatility is dependent on the values of volatility observed in the recent past?
What if they also depend on the latest returns?
In an exponentially weighted moving average model, the weights assigned to the u2 decline
exponentially as we move back through time
This leads to: V n2 OV n2 1 (1 O )u n21
Apply the recursive relationship:
Hence we have
m
V n2 (1 O ) Oi 1u n2i OmV n2 m
i 1
Variance estimate for next day (n) is given by (1-
previous variance estimate
Risk-metrics (by JP Morgan) assumes a Lambda of 0.94
Since returns are squared, their direction is not considered. Only the magnitude is considered
In EWMA, we simply need to store 2 data points: latest return & latest volatility estimate
Consider the equation: V t 1 (1 0.94) P t 0.94V t
2 2 2
In this equation, variance for time t was also an estimate. So we can substitute for it as follows:
V t21 (1 0.94) Pt2 0.94>(1 0.94) Pt21 0.94V t21 @
Example 1: On Tuesday, return on a stock was 4%. Volatility (Std. deviation) estimate for Tuesday
&t
Variance estimate for Wednesday = (1-2
Std. Dev. = sqrt (1.9%2)=1.378%
Tuesday volatility (Std. Dev.) estimate was 1%. Actual return on Tuesday was 4%. Therefore, volatility
estimate for Wednesday is estimated upwards than Tuesday i.e. 1.378% as compared to 1%.
Notice how the volatility estimate has been revised due to high return
EWMA Question
Example 1: Continuing the previous example, volatility estimate for Wednesday was 1.378%.
Assume that actual return on Wednesday was 0%. What is the variance estimate for Thursday?
Solution: Variance estimate for Thursday = (1-2 Std. Dev. = 1.34%
In very short-term like daily returns, estimated volatility is the expected return
Since latest return of 0% was lesser than estimated volatility (and estimated return) of 1.378%, volatility for
next day is revised downward from 1.378% to 1.34%
Notice the downward revision in the estimate due to lower return
,&
hZDtD
conditional variance, which weight will be applied to the return that is 4 days old?
A. 0.000
B. 0.043
C. 0.048
D. 0.950
Solution
B.
A. Incorrect. The wrong factor has been squared. The EWMA RiskMetrics model is defined as:
t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of (1-
0.95)3*0.95 = 0.00012 for r0 when t = 4.
B. Correct. The EWMA RiskMetrics model is defined as:
t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of (1-
0.95)*0.95^3 = 0.043 for r0 when t = 4.
C. Incorrect. The 0.95 has not been squared. The EWMA RiskMetrics model is defined as:
t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of (1-
0.95)*0.95 = 0.048 for r0 when t = 4.
/ddtDZD
t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of 0.95 =
0.950 for r0 when t = 4.
V t21 JV L DP t2 EV t2
Long-term average Volatility
V t21 Z DP t2 EV t2
'sL
Since the sum of all the weights is equal to 1 we get the following equation as well:
Z
VL
1 D E
Suppose a GARCH model is estimated using MLE from daily data as follows:
Solution
Solution: In the GARCH model, 12% is the weight given to latest squared return (reactive factor).
85% is the weight given to latest variance estimate (persistence factor). Therefore,
1-0.12-0.85 = 3% is weight given to long-term average Volatility.
Therefore, 3%*VL = 0.000005 i.e. VL = 0.017%
-
s^
&'Z,/-term volatility is
negative and the model becomes mean-fleeing
Which of the following GARCH models will take the shortest time to revert to its mean?
A. ht 2t-1 t-1
B. ht 2t-1 t-1
C. ht 2t-1 t-1
D. ht 2t-1 t-1
Solution
B.
A. Incorrect. The model that will take the shortest time to revert to its mean is the model with the lowest
/
B. Correct. The model that will take the shortest time to revert to its mean is the model with the lowest
/
C. Incorrect. The model that will take the shortest time to revert to its mean is the model with the lowest
/
D. Incorrect. The model that will take the shortest time to revert to its mean is the model with the lowest
/
Suppose the long-run variance rate is 0.0002 so that the long-run volatility per day is 1.4%
Solution
EduPristine www.edupristine.com
Agenda
MCS is used to model complex problems or estimate variables for where sample size is small. It is
used in finance in situation such as:
The pricing of exotic options, where an analytical pricing formula is unavailable
Determining the effect on financial markets of substantial changes in the macroeconomic
environment
Stress-testing risk management models to determine whether they generate capital requirements
sufficient to cover losses in all situations.
Cont...
In the first step, a model needs to be specified to generate the data. The model can be pure time
series or a structural model. Pure time series model are simpler whereas structural model requires a
DGP for the explanatory variables and hence is more complex.
In the second stage, estimation about the parameter of interest is made. Here various scenarios can
also be developed about the outcome of the variable.
In the last two stages, data analysis is carried out and the steps are repeated N times.
The sampling variation in Monte Carlo is estimated by the standard error estimate (Sx)
var( x)
Sx
N
It is evident from the formula that to reduce the error by a factor of 10, N must be increased by 100.
meaning, in order to improve the efficiency, the N must be significantly higher which is both time
consuming and costly. The alternative way to reduce the sampling error is to use variance reduction
techniques. Two most widely used variance reduction techniques are
1. Antithetic variates
2. Control variates
Antithetic variates
Monte Carlo requires a lot of repetition to adequately cover the entire probability space. By their
very nature, the values of the random draws are random, and so after a given number of
replications, it may be the case that not the whole range of possible outcomes has actually occurred.
What is really required is for successive replications to cover different parts of the probability space
that is, for the random draws from different replications to generate outcomes that span the entire
spectrum of possibilities. This may take a long time to achieve naturally. Hence, a an alternative
approach , the antithetic variate technique is used which reduces the sampling error by returning a
simulation using a complement set of original set of random variables.
For example, if the driving stochastic force is a set of TN(0, 1) draws, denoted u1, for each
replication, an additional replication with errors given by u1 is also used. It can be shown that the
Monte Carlo standard error is reduced when antithetic variates are used.
Suppose the average value of parameter of interest across two set of Monte carlo replication is given
by
x ( x1 x 2 ) / 2
Where x1 and x2 are the average output parameter values for simulation sets 1 & 2, respectively.
The variance of x is given by,
var( x)
1
var( x1 ) var( x2 ) 2 cov( x1 , x2 )
4
If no antithetic variates are used, the two sets of Monte carlo replications are independent and there
co-variance will be zero leading to variance of :
var( x)
1
var( x1 ) var( x2 )
4
The use of antithetic variate results in negative covariance between the original random draws and
their complements which will further reduce the Monte Carlo sampling error.
Control variates
The application of control variates involves employing a variable similar to that used in the
simulation, but whose properties are known prior to the simulation. Denote the variable whose
properties are known by y, and that whose properties are under simulation by x. The simulation is
conducted on x and also on y, with the same sets of random number draws being employed in both
cases. Denoting the simulation estimates of x and y by x and y, respectively, a new estimate of x
can be derived from:
x* y ( x y )
The new x* variable estimate will have a smaller sampling error than the original x variable if the
control statistic and statistic of interest are highly co-related.
The variance of y would be zero as it is control variable and its properties are known. So, the
variance would be:261
var( y ) 2 cov( x , y ) 0
The same can be simplified as
1
cov( x , y ) ! var( y )
2
By dividing both the side of this inequality by product of Standard deviations , co-relation can be
obtained:
1 var( y )
corr ( x , y ) !
2 var( x )
Cont...
For pricing an arithmetic Asian option using simulation. An arithmetic Asian option is one whose
payoff depends on the arithmetic average value of the underlying asset over the lifetime of the
averaging; at the time of writing, an analytical (closed-form) model is not yet available for pricing
such options. In this context, a control variate price could be obtained by finding the price via
simulation of a similar derivative whose value is known analytically e.g. a vanilla European option.
Thus, the Asian and vanilla options would be priced using simulation with the simulated price given
by P(A) and P(BS), respectively. The price of the vanilla option, P(BS) is also calculated using an
analytical formula, such as BlackScholes. The new estimate of the Asian option price, P*A, would
then be given by
Bootstrapping
Unlike simulation, bootstrapping approach draws random return data from a historical data. It
involves sampling repeatedly with replacements from the actual data.
Suppose a sample of data, y= y1, y2, . . . , yT are available and it is desired to estimate some
parameter . An approximation to the statistical properties of d can be obtained by studying a
sample of bootstrap estimators. This is done by taking N samples of size T with replacement from y
and re-calculating with each new sample. A series of estimates is then obtained, and their
distribution can be considered.
The advantage of bootstrapping: No assumption are made regarding the true distribution of the
parameter estimate,
How bootstrapping is used:
1. Generate a sample of size T from the original data by sampling with replacement from the whole
rows taken together (that is, if observation 25 is selected, take y25 and all values of the
explanatory variables for observation 25).
2. Calculate , the coefficient matrix for this bootstrap sample.
3. Go back to stage 1 and generate another sample of size T. Repeat these stages a total of N times.
A set of N coefficient vectors, , will thus be obtained and in general they will all be different, so
that a distribution of estimates for each coefficient will result.
Non-Independent data: Bootstarp assumes that the data are independent of one another. However,
if there is autocorrelation in the data, then it is not useful. To overcome the same, a method called
moving block bootstap can be used wherein blocks of data are examined at one time in order to
preserve the original data dependency.
Recursive process for generating the random number will require the user to specify an initial value
to get the process started. The choice of this value will, undesirably, affect the properties of the
generated series. This effect will be strongest for y1, y2, . . . , but will gradually die away. For
example, if a set of random draws is used to construct a time series that follows a GARCH process,
early observations on this series will behave less like the GARCH process required than subsequent
data points. Consequently, a good simulation design will allow for this phenomenon by generating
more data than are required and then dropping the first few observations. For example, if 1,000
observations are required, 1,200 observations might be generated, with observations 1 to 200
subsequently deleted and 201 to 1,200 used to conduct the analysis.
These computer-generated random number draws are known as pseudo-random numbers, since
they are in fact not random at all, but entirely deterministic, since they have been derived from an
exact formula! By carefully choosing the values of the user-adjustable parameters, it is possible to
get the pseudo-random number generator to meet all the statistical properties of true random
numbers. Eventually, the random number sequences will start to repeat, but this should take a long
time to happen.
Computationally intensive
Results are not precise: The simulation process is based on some assumption about the
parameters. However, if there is any error in the assumption in data generating process, it can lead
to in accurate results. For example, in the context of option pricing, the option valuations obtained
from a simulation will not be accurate if the data generating process assumed normally distributed
errors while the actual underlying returns series is fat-tailed.
Results are hard to replicate: Practically, MCS is not conducted documenting each sequence of
random draw and hence results will be specific with each draw. If number of replication is small, a
repeat of the experiment would involve different sets of random draws and therefore would be
likely to yield different results.
Results are experiment specific: Simulation process in finance is based on specific assumptions
related to data generating process and set of equations. If the assumptions are changed in data
generating process or set of equation, the results can be significantly different.
Exercise
1. A risk manager was trying to understand on how to reduce the sampling error in MCS. His senior
told him about the two methods and asked him to use the most suitable method for his problem.
After going through the research provided by the seniors, the manager made following comment:
A. The manager concluded that the antithetic method will reduce the error due to perfectly negative co-
relation between the parameter and its complement while using Asian option valuation.
B. Under control variate technique, the controlled variable has no sampling error.
Which of the following option is correct?
1. A only
2. B only
3. A & B both
4. Neither A nor B
2. One can not change the assumptions of data generating process or set of equation while
implementing MCS . If the above condition is violated, which one of the limitation of MCS it will
fall in?
1. High computation cost
2. Results are imprecise
3. Results are difficult to replicate
4. Results are experiment specific
2. Option 4
dz
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
FMP-I
Forward and Futures
EduPristine www.edupristine.com
AGENDA
Introduction Derivatives
Derivatives are financial instruments which derive their value from an underlying asset and some
other variables such as interest rates, volatilities etc.
Futures, forwards, options and swaps are some of the most common examples of derivatives.
The underlying asset: It is a more basic financial instrument. Example: stocks, bonds, interest rate,
commodity etc.
Asset types
Financial Assets: Equity, Debt securities.
Commodities: Gold, Copper, Crude Oil.
Real Estate
Lets take an example of a financial asset (stock)
We can buy the stock through the broker by paying the stock price.
We can either hold the bought asset or sell it at the current market price.
During the holding period of the stock, the dividends received goes to your pocket as the income from
the asset.
After selling the asset, we earn a profit or loss on the asset, depending on the selling price of the
asset (stock).
Purpose of Assets
Investment Asset and Consumption Asset.
Market Maker
An individual or an institution which keeps an inventory of financial instruments or commodities who could
be asked for trading those assets. The individual or the institution then quotes a bid and an ask price on the
option.
Futures Contracts: Agreement to buy or sell an asset for a certain price at a certain time. A futures
contract is traded on an exchange.
Forward Contracts: Forward contracts are similar to futures except that they trade in the over-the-
counter market.
Notation for Valuing Futures and Forward Contracts
S0: Spot price of the asset underlying today.
F0: Futures or forward price today.
T: Time until delivery date (in years).
R: Risk-free interest rate per annum, expressed in continuous compounding, for maturity T.
Payoff of forwards and futures:
Long Short
P P
X S X S
In both Forward and Futures contracts there is an obligation to buy or sell an asset
Options
(to be covered in detail in later slides)
100 100
Asset Asset
Price Price
100 100
Hedgers
Uses derivatives to hedge the risks they face from volatility in the asset prices.
Example: a company is getting a cash inflow in 3 months time in a foreign currency. It hedges its currency risk
by taking a short position in a currency forward at a particular price.
Speculators
Uses derivatives to bet on a particular direction of movement of the asset price.
If a speculator believes that the spot rate of a foreign currency will be higher in 3 months than its present 3
month forward rate, he goes long on the forward. After 3 months if he is correct, he receives foreign currency
at lower rate and immediately resells it at the higher spot rate.
Arbitrageurs
Take offsetting positions in 2 or more instruments to lock a profit.
Suppose that:
The spot price of gold is US$390.
The quoted 1-year forward price of gold is US$425.
The 1-year US$ interest rate is 5% per annum.
No income or storage costs for gold. Is there an arbitrage opportunity?
&Z^d&
Arbitrage = Buy Low Sell High at no risk
An investor bought 1000 shares of ABC company each priced at $50. The initial margin
requirement were 60%.and the maintenance margin requirement is 25%. At what price would the
investor be getting a margin call?
The investor gets a call when he/she loses 30,000 12,500 = 17,500
Price of share after this loss = 50 17.5 = $32.50
Hence the investor will get the margin call when the price falls to $32.50
What would be the variation margin if the stock price reduced to $10 from $50?
When stock goes down to $10, the loss = 1000 x (50 10) = 40,000
Hence margin account becomes 30,000 40,000 = -10,000
Hence the investor will need to pay [(30,000 (-10,000)] = 40,000 as variation margin
Day Initial price Settlement price Daily Gain/Loss Cumulative Gain/Loss Margin account Margin call
1-Mar $100.00 $10,000.00
1-Mar $100.00 $101.00 $1.00 $200.00 $10,200.00
2-Mar $101.00 $101.50 $0.50 $100.00 $10,300.00
3-Mar $101.50 $98.50 -$3.00 -$600.00 $9,700.00
4-Mar $98.50 $96.50 -$2.00 -$400.00 $9,300.00
5-Mar $96.50 $92.00 -$4.50 -$900.00 $8,400.00 $1,600.00
6-Mar $92.00 $91.00 -$1.00 -$200.00 $9,800.00
7-Mar $91.00 $90.20 -$0.80 -$160.00 $9,640.00
8-Mar $90.20 $93.80 $3.60 $720.00 $10,360.00
9-Mar $93.80 $80.50 -$13.30 -$2,660.00 $7,700.00 $2,300.00
10-Mar $80.50 $90.90 $10.40 $2,080.00 $12,080.00
Very few futures contract lead to a delivery of the underlying asset. Most are closed out early
The period of delivery is decided by the exchange, but the exact date of delivery is specified by the
short contract holder
Notification to deliver is given by the broker to the clearing house. The number of contracts and
the specifics of the delivery (what grade, type, quality, location, etc) is mentioned
Exchange identifies a party with a long position to accept delivery. Typically one with the oldest
outstanding long position
Party with the long position has to accept the delivery
Whole delivery process from the issuance of intention to deliver to delivery takes 23 days
Cash settlement is also possible. Settlement price is the spot price of the underlying on the day
opening/day close
Types of orders
Precious Metals Settlement price: this price is used for margin call
Gold Comex (100 Troy oz; $/troy oz) calculations at end of trading day
Sett Days
High Low
Vol 0 int Change: -$8.5 per troy ounce. Hence, total of
price chge 000s 000s 100*$8.5=$850 reduction in margin account balance
Dec 738.7 -8.5 748.5 736.7 111.8 300.3 Open interest: total number of contracts outstanding
Feb 745.0 -8.6 754.8 743.5 6.00 31.86 with the exchange. It is the total number of long
Total 123.7 449.5 positions / the total number of short positions. It is
one trading day older than the prices day
Time Trading Activity Open Interest
Jan 1 A buys 1 option and B sells 1 option contract 1
Jan 2 C buys 5 option and D sells 5 option contract 6
Jan 3 A sells his 1 option and D buys 1 options contract 5
Jan 4 E buys 5 options from C who sells 5 options contracts 5
On January 1, A buys an option, which leaves an open interest and also creates trading volume of 1
On January 2, C and D create trading volume of 5 and there are also five more options left open
On January 3, A and D take offsetting positions, open interest is reduced by 1 and trading volume is 1
On January 4, E simply replaces C and open interest does not change, trading volume increased by 5
Generally, there is a high correlation in price movements between the futures market and the
cash market.
Futures position acts as a substitute for later cash transaction.
Hedgers generally take equal and opposite position in cash and futures.
Hedging and shareholders: Shareholders can hedge the risk themselves. Companies dont need
to. But do the shareholders have as much information as the companies? What about transaction
costs and commissions? Companies carry out high volume transactions hence cost of hedge
is lower.
Hedging and competitors: What if the price of hen food was reduced as the hen producers union
pressed the suppliers to reduce their prices. HPs profits would rise as he had locked his selling
price and the raw material prices went down. For others the change in profits would be 0. What if
the raw material prices went up for some reason and the union decided to raise their selling
prices in the market proportionately. HPs profit would reduce while others profit remain the
same. Hedging actually is causing fluctuation in profits!!
Not always does the futures contract date be the same as the
date the asset is to be bought or sold. Future
What if the farmer didnt know when his corn produce would be Price
b1 = S1 - F1 and b2 = S2 - F2
Future
Farmer pay off when he sells his corn: S2 F1 - F2 or F1 b2 Price
In a typical transaction, F1 is known but b2 is not known at time
t1 b2 is the basis Time
Expect 2 questions directly based on the 4 factors that incorporate basis risk
Imagine a stack-and-roll hedge of monthly commodity deliveries that you continue for the next
five years. Assume the hedge ratio is adjusted to take into effect the mistiming of cash flows but is
not adjusted for the basis risk of the hedge. In which of the following situations is your calendar
basis risk likely to be greatest? (FRM 2008 Sample Paper)
A. Stack and roll in the front month in oil futures
B. Stack and roll in the 12-month contract in natural gas futures
C. Stack and roll in the 3-year contract in gold futures
D. All four situations will have the same basis risk
Solution
A.
The oil term structure is highly volatile at the short end, making a front-month stack-and roll hedge heavily
exposed to basis fluctuations. In natural gas, much of the movement occurs at the front end, so the 12-month
contract wont move much. In gold, the term structure rarely moves and wont begin to compare with oil and
gas.
In such cases the proportion of the exposure that should optimally be cross hedged
V
Optimal Hedge Ratio: h
U G S
V GF
Where
S is the standard deviation of ^ the change in the spot price during the hedging period.
F is the standard deviation of & the change in the futures price during the hedging period.
is the coefficient of correlation between ^ and &.
In the case of a stock index the similar logic follows. If P is the value of the portfolio of stocks held
by an investor and A is the current value of the stocks lying under one futures contract then the
optimal hedge ratio, N*, should be equal to P / A.
In practical cases investors dont typically have portfolios that trace the index. Hence the concept
of comes into play. Beta is a measure of a stock's volatility in relation to the market.
In order to change the beta of the portfolio to we need to long or short the (N*) number
of contracts depending on the sign of (N*)
P
N*
A
P
N * ( E * -E )
A
Negative sign of (N*) indicates shorting the contracts
If you have a portfolio of $500,000 which mirrors S&P 500. Each S&P 500 contract is $250 times
the index when the index is at 500. Calculate the number of contracts to be hedged?
Solution
The current value of the S&P 500 index is 1,457, and each S&P futures contract is for delivery of
US$250 times the index. A long-only equity portfolio with market value of US$300,100,000 has
beta of 1.1. To reduce the portfolio beta to 0.75, how many S&P futures contract should you sell?
(FRM Sample Paper 2009)
Solution
Investment assets
Assets held for investment purposes by significant numbers of investors.
(examples: stocks, bonds, gold, silver)
Consumption assets:
Assets held primarily for consumption (examples: copper, oil and pork bellies)
Gambling Short Selling an example
Short selling involves selling securities that are not owned
Suppose an investor short sells 500 IBM shares, the broker will borrow the securities from another client and sells
them in the market in the usual way
At some stage the investor will close the position by purchasing 500 IBM shares. The investor takes the profit if the
stock prices have declined , else vice versa
Short Squeezed: If anytime the broker runs out of shares to borrow, the investor is short squeezed and forced to
close his position immediately
Forwards Futures
Not traded on exchanges Traded on exchanges
Are private agreements between two parties Standard contracts
and are not as rigid in their stated terms and Clearing house and daily mark to market
conditions reduces credit risk
Credit risk is high Settlement can occur over a range of dates
High customization Usually closed out before maturity and hardly
Settlement at the end of contract and on a any deliveries happen
specific date
Mostly used by hedgers that want to remove the
volatility of the underlying, hence delivery/ cash
settlement usually takes place
At the time of entering into a forward contract, long or short, the value of the forward is zero
This is because the delivery price (K) of the asset and the forward price today (F0) remains the
same
The value of the forward is basically the present value of the difference in the delivery price and
the forward price
Value of a long forward, f, is given by the PV of the pay off at time T:
= (F0 K )erT
K is fixed in the contract, while F0 keeps changing on an everyday basis
Forward and futures prices are usually assumed to be the same. When interest rates are uncertain,
they are slightly different in theory
A strong positive correlation between interest rates and the asset price implies the futures price is
slightly higher than the forward price
A strong negative correlation implies the reverse
Formula to remember:
If Spot rate is given in USD/INR terms then take American Risk-free rate as the first rate
In other words, individual who is interested in USD/INR rates would be an American (Indian will always think
in Rupees not dollars!!), which implies foreign currency (rf) in his case would be rINR
The forward rate of a 3-month EUR/USD foreign exchange contract is 1.1565 USD per EUR. USD
LIBOR is 4% and EUR LIBOR is 2%. The spot USD per EUR exchange rate is?
Solution
F0 = S0 e(r-rf)t
1.1565 e-(.04- .02).25 = 1.1507
Assume that the current 1-year forward exchange rate is 1.200 USD per EUR. An American bank
pays 2.4% annual interest rate on a 1-year deposit and a 4.0% annual interest rate on a 3-year USD
deposit. A European bank pays a 1.5% annual interest rate for a 1-year deposit and a 2.0% annual
interest rate for a 3-year EUR deposit. The forward exchange rate of USD per EUR for exchange
three years from today is closest to:
Solution
The two-year risk-free rate in the United Kingdom is 8% per annum, continuously compounded.
The two-year risk-free rate in France is 5% per annum, continuously compounded. The current
French Franc to the GBP currency exchange rate is 1GBP = 0.75 French Franc.
What is the two-year forward price of one unit of the GBP in terms of the French Franc so that no
arbitrage opportunity exists?
A. 0.578
B. 0.706
C. 0.796
D. 0.973
Solution
A bank has a USD 50,000,000 portfolio available for investing. The cost of funds for the USD
50,000,000 is 4.5%. The bank lends 50% of the assets to domestic customers at an average loan
rate of 6.25%. The rest of the portfolio is lent to UK clients at 7%. The current exchange rate is
USD1.642/GBP. At the same time, the bank sells a forward contract equal to the expected receipts
one year from now. The forward rate is USD1.58/GBP. The weighted average return to the bank on
its investments is closest to?
Solution
The return from UK customers, $25,000,000/1.642 = GBP 15,225,335* 1.07 = GBP 16,291,108
The bank sells a forward contract: GBP 16,291,108*1.58 = USD 25,739,951
Earnings (USD 25,739,951 25,000,000) / 25,000,000 = 2.96%
t
Solution
The spot is quoted in terms of Swiss Francs per USD. To convert this into USD per Swiss Franc, we
get: 1/1.3680 = 0.7310. The theoretical futures price = 0.7310 * exp((0.0105 0.0035) * 0.25) =
0.7323. Therefore, the quoted futures price is too high. Thus, one should sell the overvalued CHF
futures contract.
In order to arbitrage, one would do the following:
Borrow USD 0.7310 * exp((-0.0035)*0.25) = USD 0.7304 for 3 months
Buy spot exp((-0.0035)*0.25) = CHF0.9991, invest at 0.35% for 3 months
Short a futures contract on CHF1
At maturity,
Pay back 0.7304 * exp((0.0105) * 0.25) = USD 0.7323
Receive 0.9991 * exp((0.0035) * 0.25) = CHF 1
Delivers CHF 1 on the futures contract, receives USD 0.7350
An arbitrage profit of USD0.7350 USD0.7323 = USD 0.0027 would be realized in 3 months time
The current spot price for cotton is $0.325 per pound. The annual risk-free rate is 3.0%, and the
cost to store and insure cotton is $0.002 per pound per month. A 3 month futures contract for
cotton is trading at $0.3368 per pound. Is there an arbitrage opportunity available, and if so, how
should an investor take advantage of it?
Solution
The Forward price = 0.325e0.03*0.25 (0.002 0.002*1.0025 0.002*1.0025^2) = 0.3335; Yes; the
investor should sell the futures contract, borrow at the risk-free rate, and buy the
spot asset.
The spot rate for a commodity is $19. The annual lease rate for the commodity is 5%. The
appropriate continuously compounding annual risk-free rate is 6.5%. What is the 3-month
commodity forward price?
Solution
F0 = S0 e(r- G)t = $19.07
Futures Prices
In a treasury futures contract the price of the bond is not easy to deliver because the cheapest to
deliver bond is not known
However, if the cheapest to deliver bond and its delivery date is known we can call upon the
equation which considers discreet payouts from an underlying and can be given as below:
F0 = (S0-I) ert
Eurodollar Futures:
A Eurodollar is a dollar deposited in a foreign bank/US bank outside the United States
Eurodollar futures are futures on the 3-month Eurodollar deposit rate (same as 3-month LIBOR
rate)
Long position => agrees to give a loan at the determined price
One contract is on the rate earned on $1 million
A change of one basis point or 0.01 in a Eurodollar futures quote corresponds to a contract price
change of $25 (1mm * 0.01% * 90/360)
When it expires (on the third Wednesday of the delivery month), final settlement price is 100
minus actual three month deposit rate.
Contract Price = 10,000*[100 0.25*( 100 Q)]
Q = Quoted Price
Question
Suppose you buy (take a long position in) a Eurodollar futures contract on November -1
The contract expires on December-21 Date Quote
The prices are as shown Nov 1 97.12
How much do you gain or lose Nov 2 97.23
On the first day Nov 3 96.98
On the second day
Over the whole time until expiration?
Dec 21 97.42
1
Forward rate futures rate V 2T1T2
2
V is the standard deviation of the change in the short term interest rate in 1 year
When the futures contracts have substantial time to maturity then the futures prices are different
from the expected future spot prices
When futures prices are greater than the expected future spot prices then the scenario is termed
as contango
When futures prices are lower than the expected future spot prices then the scenario is termed as
normal backwardation
Normal futures curve: When futures prices are greater for greater maturity
Inverted futures curve: When futures prices are lower for greater maturity. (example: orange juice,
because its value depreciates with time)
Commodity Spreads
Commodity Spread: is a result of a commodity that is used in the production process. Lets take an
example of mustard seeds which can be used to prepare mustard oil which sells at a higher price
than mustard seeds. This difference of prices between the raw and processed commodity is the
commodity spread
Commonly used Commodity spreads:
Commodity Spreads
Optimal Hedge Ratio: Optimal number of contracts: Future Price: Interest Rate Parity:
V S USD e ( rUSD rINR )T
h
U GS N* ( E * -E )
P F0 = S0erT FUSD
V GF A F0 = S0 (1+r )T INR INR
dz
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
FMP-II
Interest Rates
EduPristine www.edupristine.com
Agenda
This reading also covers the following readings from Valuation and Risk Models
Spot, forward and Par Rates
Returns, Spreads and Yields
Interest rate is the amount of money a borrower promises to pay to the lender over and above the
principal amount
Treasury Rates: This is the rate an investor receives when he invests in Treasury bills and Treasury bonds.
Treasury bills are short term while Treasury bonds are longer term (> 1 year)
Corporate bond rates: These are rates on long term bonds issued by a corporate
LIBOR: This is the London Interbank Offer Rate (LIBOR) and the rate at which banks make a large wholesale
deposit or loan with/to another bank
1 month, 3 months, 6 months and 12 month LIBORs
Opportunity cost for AA rated banks
Not entirely risk free
Repo rates and Reverse Repo: Repo rate is the rate at which banks borrow money from the central bank.
Reverse Repo rate is the rate at which the central bank borrows money from banks
A Repurchase agreement (also known as a repo or Sale and Repurchase Agreement) allows a borrower to use
a financial security as collateral for a cash loan at a fixed rate of interest
A repo is equivalent to a cash transaction combined with a forward contract
There are many ways to calculate interest rates annual, semi annual, quarterly, continuously
compounding and so on
Each rate can be expressed in the form of another rate. For example an interest rate of 10%
compounded semi-annually would fetch (1 10% / 2) * (1 10% / 2) = 1.1025 (remember
6months rate is 10% / 2) on $1 after one year. This is equivalent of 10.25% annual rate
Amount compounded annually would be given by:
A = P (1+ r)t
A terminal amount
P principal amount
r annual rate of interest
t number of years for which the principal is invested
If amount compounded n times a year then:
A = P ( 1 r/n )nt
When n then we call it continuous compounding:
A = Pert (this formula is derived using limits and continuity)
Question
If the interest rate is 10% per annum compounded continuously, then what is the effective annual
interest rate?
Question
If the interest rate is 10% per annum compounded semi-annually then what is the equivalent
continuously compounded interest rate.
2 = 1e(rx1)
=> 1.1025 = er
=> r = 0.09758 = 9.758%
Introduction Bonds
A bond is a debt security usually issued by a company or the government to raise funds
Example: A company ABC issues bonds of worth $100. An investor X buys the bond by paying
$100 to the company ABC. ABC promises to repay the money back to X after 5 years and also pay
5% of the $100 principle every year, semi-annually
In the above example:
Face Value: $100
Coupon rate: 5%
Time to maturity: 5 years
W
C C C C
Bonds are either zero coupon bonds (having no interest payments) or coupon bonds
(with periodic interest payments)
The price of a bond is the present value of all the coupon payment and the final principal payment
received at the end of its life
1
T
1
B Ce Pe
rt rT
(1 YTM) n
1
t 1 B Iu Fu
B the bond price YTM (1 YTM) n
C coupon payment
r zero interest rate at time t
P bond principal
T time to maturity
The yield of a bond is the discount rate (applied to all future cash flows) at which the present
value of the bond is equal to its market price
z to Maturity = Investors Required Rate of Return
The par yield is the coupon rate at which the present value of the cash flows equal to the par
value (principal value) of the bond
If we are looking at a semi-annual 5 year coupon bond with a par value of $100 then the coupon
payment would be solved using the following equation:
5
100 (C / 2)e
t 1
rt
100e 5 r
Question
Or
z 1 2 3 4 5 6 7 8 9 10
z 12%
Coupon payments 7 7 7 7 7 7 7 7 7 7
Principal payment 100
PV factor 0.892857 0.797193878 0.711780248 0.635518078 0.567427 0.506631 0.452349 0.403883 0.36061 0.321973
Total PVs 6.25 5.580357143 4.982461735 4.448626549 3.971988 3.546418 3.166445 2.827183 2.52427 34.45114
Bond price 71.74888
z 1 2 3 4 5 6 7 8 9 10
z 13%
Coupon payments 7 7 7 7 7 7 7 7 7 7
Principal payment 100
PV factor 0.884956 0.783146683 0.693050162 0.613318728 0.54276 0.480319 0.425061 0.37616 0.332885 0.294588
Total PVs 6.19469 5.482026784 4.851351136 4.293231094 3.79932 3.36223 2.975425 2.633119 2.330194 31.52095
Bond price 67.44254
In the case of treasury rates there are some key facts to know:
Treasury bills are issued at a discount from face value and are paid at their par (face amount) at maturity.
The purchase price is expressed as a price per hundred dollars
Bills are sold at a discount. The discount rate is determined at auction
Bills pay interest only at maturity. The interest is equal to the face value minus the purchase price
Bills are sold in increments of $100. The minimum purchase is $100
Boot Strap Method to determine zero rates
Consider the bond prices of Treasury bonds given below in column 4. Calculate the continuously
compounded zero rates for 6 months, 12 months, 18 months and 24months
Continuously
Bond Principal Time to Maturity Annual Coupon Bond Price Compounded 0-
rate
100 0.5 10 99.5 10.76
100 1.0 10 98.4 11.43
100 1.5 10 96.5 12.31
100 2.0 10 94.3 13.01
A forward rate agreement (FRA) is an over the counter agreement where the forward interest rate,
Ft1,t2 ,is fixed for a certain principal between times T1 and T2
The payer of the fixed interest rate is also known as the borrower or the buyer. The buyer hedges against
the risk of rising interest rates, while the seller hedges against the risk of falling interest rates
Payment to the long at settlement = Notional Principal X (Rate at settlement FRA Rate) (days/360)
----------------------------------------------------------
Z
Question
Duration
Duration it is the measure of how long on an average the holder of the bond has to wait before he
receives his payments on the bond
A coupon paying bonds duration would be lower than n as the holder gets some of his payments
in the form of coupons before n years
Macaulays duration: is the weighted average of the times when the payments are made. And the
weights are a ratio of the coupon paid at time t to the present bond price
n
t *C n*M
(1 y)
t 1
(1 y ) n
t
Macaluay Duration
Current bond price
Where:
t = Respective time period
C = Periodic coupon payment
y = Periodic yield
n = Total no of periods
M = Maturity value
Macaulay duration is also used to measure how sensitive a bond or a bond portfolio's price is to
changes in
interest rates
A bonds interest rate risk is affected by:
Yield to maturity
Term to maturity
Size of coupon
From Macaulays equation we get a key relationship:
'B
D'Y
B
In the case of a continuously compounded yield the duration used is modified duration given as:
Macaulay Duration
D*
r
1
n
Consider a bond trading at 96.54 with duration of 4.5 years. In this case:
- 96.54* 4.5 -434.43
/) in the yield then the bond price would change by:
-434.43 * ( 0.001)
-0.43443
Hence, B = 96.54 0.43443 = 96.10
Convexity
Tangent
Y* Yield
The convexity of the price / YTM graph reveals two important insights:
The price rise due to a fall in YTM is greater than the price decline due to a rise in YTM, given an identical
change in the YTM
For a given change in YTM, bond prices will change more when interest rates are low than when they are high
To make the convexity of a semi-annual bond comparable to that of an annual bond, we can divide
the convexity by 4
In general, to convert convexity to an annual figure, divide by m2, where m is the number of
payments per year
We can approximate the change in a bonds price for a given change in yield by using duration and
convexity:
'VB D Mod
u 'i u VB 0.5 u C u VB u 'i
2
If yields rise by 1% per period, then by what price will the bond fall by? Assume C = 16.65.
Solution
Note:
The 1st term -D x x P , can either be a positive or a negative amount, indicating an increase or
decrease in price respectively. Depending on the sign of .
Term structures
The term structure of interest rates is graphed as though each coupon payment of a non callable
fixed-income security were a zero-coupon bond that matures on the coupon payment date
The yield curve describes the yield differential among treasury issues of differing maturities
The Yield Curve is the graph created by putting term to maturity on the X axis, YTM on the Y axis
and then plotting the yield at each maturity
Upward sloping: This is the most persistent shape historically when short-term interest rates and inflation
are low
Downward sloping (Declining): This occurs at peaks in the short-term interest rate cycle, when inflation is
expected to decrease in the future
Flat: This shape is evident during periods of interest rates transitions
Humped: This occurs when rates are transitioning or perhaps market participants are attracted in large
numbers to particular maturity segment of the market, thereby creating the hump
16
14
12
z^
10
Percent
0
1 mth 3 mths 6 mths 1 yr 2yrs 5 yrs 7 yrs 10 yrs 30 yrs
Questions
Discuss the yield curve below and the economic impacts it conveys:
Yield
Maturity
What is the 1.5 year par yield of a $100 bond when the zero rates (Continuous Compounded) on
6months, 12 months and 18 months are 4%, 4.5% and 5%?
Sol: Let it be C
100 = C*e-0.5*.04 -1*.045 -1.5*.05
C = 7.225651/2.863939
C = 2.523
Coupon = 2.523*2 = 5.05%
Day count defines the way in which interest is accrued over time. Day count conventions normally
used in US are:
Actual / actual treasury bonds
30 / 360 corporate bonds
Actual/360 money market instruments
The interest earned between two dates
Treasury bonds
The prices for treasury bonds are quoted in dollars and 1/32nd of a dollar
$8227 is equivalent to $82.84375
Cash price / dirty price is the price at which the investor buys a bond from the market
Cash price = Quoted price + accrued interest
Accrued interest is the interest which the nearest coupon that is due generates
Conversion price
When a bond is delivered the party with the short position, the amount transacted is:
Quoted futures price + accrued interest
Where, Quoted futures price = settlement price conversion factor
The conversion factor is equal to the quoted price the bond would pay per dollar or principal on
the first day of the delivery month on the assumption that the interest rate for all maturities
equals 6% per annum (with semi annual compounding)
Question
The last coupon payment of $10 was paid on a treasury bond on June 19, 2009. The next coupon
is due on December 19, 2009 and we are currently on September 1, 2009. If the quoted price is
$8227 then the cash price would be?
At any given time during the delivery month there are many bonds that can be delivered in the
CBOT futures contract
The party with the short position can chose to deliver the cheapest bond when it comes to
delivery, hence he would chose the cheapest to deliver bond
Net pay out for delivery (he has to buy a bond and deliver it):
Quoted bond price
Consider an example in the table below where the short position holder has 3 options for delivery.
His cheapest to deliver bond is Bond 2
Cheapest to Deliver Bond (All figures in $)
Settlement Future Price: 94.23
Bond Quoted Bond Price Conversion Factor
1 99.6 1.033
2 135.67 1.432
3 122.45 1.257
DV01
DV01: The price value of basis point (PVBP) or dollar value of basis point (DV01) change is the
absolute change in the bond price from one basis point change in yield.
DV01 = ~ price at YTM0 price at YTM1 ~
YTM0 = the initial YTM
YTM1 = the YTM 1 basis point above or below YTM0 (YTM1 = YTM0 0.0001)
DVBP [ D * u P0 ] u 0.0001
Considering a situation where an asset that is interest rate dependant is hedged using an interest
rate futures contract
In such cases the number of contracts to hedge is given by the equation below:
PDP
N*
FC DF
When hedges are constructed using interest rates it is important to note that interest rate and
futures prices move in opposite directions . So if one is expecting to lose money when the interest
rate falls, one should long futures contracts so that they can hedge their losses by gains in futures
prices
An investor has invested $10m in government bonds and is expecting the interest rates to rise in
the next 6 months so he decides to hedge himself by interest rate futures. It is currently June and
he decides to use the December T-bond futures contract for the hedge. If the current futures price
is 97.2345 and the duration of the portfolio of government bonds at the end of 6 months is 7.1
years. The duration of the cheapest to deliver T-bond in December is given as 9.121 years. What
position should the investor take in the futures contract? How many futures contract should long /
short for the hedge if each contract is for the delivery of $100,000 face value?
Solution
10,000,000 7.1
u 80
(97.2345 u100,000 / 100) 9.121
Negative Convexity
Callable bonds exhibit negative convexity when yields fall below certain level.
At lower yield, there is incentive for the issuer to call the bond.
Price curve of the bond bends away from the normal curve thereby exhibiting negative convexity.
In barbell strategy, investor uses the bonds of short and long maturities and does not invest in the
bonds of intermediate maturity.
In bullet strategy, investor uses the bonds concentrated in intermediate maturity range.
In volatile rate environment, barbell strategy is preferred over bullet strategy.
Single factor approach assumes that all the future rate changes are driven by single factor.
The same change in interest rate is assumed for the entire yield curve.
In practice, change in short term interest rate might be different from the change in long term
interest rate.
Same hedging instrument cannot be used for hedging the change in short term interest rate and
long term interest rate.
Key Rates are the rates selected at key point on the yield curve. These are usually 2, 5, 10 and 30
year rates.
Key rate exposures hedge risk by using rates from a small number of available liquid bonds.
Partial 01 is used to measure the risk of the bond or swap portfolio in terms of liquid money
markets and swap instruments.
Forward 01 is used to measure the risk of the bond or swap portfolio in terms of shifts in the
forward rates.
Key Rate Shift technique is a approach to nonparallel shift in the yield curve.
This technique allows to determine changes in all the rates due to the changes in key rates.
Choice has to be made as to which key rates shifts and how the key rate movement relate to prior
or subsequent maturity key rates.
Key Rate 01 measures the dollar change in the value of the bond for every basis point shift in the
key rate
Key Rate 01 = (-1/10,000) (Change in Bond Value/0.01%)
Key rate duration provides the approximate percentage change in the value of the bond
Key Rate Duration = (-1/BV) (Change in Bond Value/Change in Key rate)
Agenda
Introduction to Options
What are Options
Intrinsic Value of Options
Returns to Option buyers and sellers
Put Call Parity
Bounds and Option Values
Determinants of Option Values
Some special cases
Summary
Options are contracts that give its buyer the right to buy or sell a particular asset
In future
At a pre-decided price (i.e. exercise or strike price)
Without any obligations
The seller of the option collects a payment (Premium) from the buyer for providing the option
Types of options:
Call or Put Options
Call Option: Gives option holder the right to buy the asset at an agreed price
Put Options: Gives option holder the right to sell the asset at an agreed price
European or American Options
European options: Are those that can only be exercised on expiration
American options: May be exercised on any trading day on or before expiration
Question
C.
A. European options cannot be exercised early
B. Small dividends will not make much of a difference in the price fall in the stock
C. A deep in the money put option should always be exercised early because it is likely that the stock might
recover from the fall
D. Though this might be profitable if the stock prices significantly fall after the ex-dividend date but the third
option is likely to provide more profit
Question
Assuming the stock price and all other variables remain the same what will be the impact of an
increase in the risk-free interest rate on the price of an American put option?
A. No impact
B. Negative
C. Positive
D. Cannot be determined
B. (Negative)
Intrinsic value: is the maximum of zero and the value of the option if the option were
exercised immediately
At the money:
When the price of the underlying is the same as the strike price of the option, the option is termed at the money
and exercising it carries a nil pay-off
In the money:
When the price of the underlying is greater than the strike price carried by a call option, the call option is termed in
the money, as exercising it results in a positive pay off
When the price of the underlying is less than the strike price carried by a put option, the put option is termed in
the money, as exercising it results in a positive pay off
Out of the money:
When the price of the underlying is less than the strike price carried by a call option, the call option is termed out
of the money, as exercising it will result in a nil pay off
When the price of the underlying is greater than the strike price carried by a put option, the put option is termed
out of the money, as exercising it will result in a nil pay off
3 0 2 5 4
4 0 1 5
5 0 0 5
2
6 1 0 5
7 2 0 5
8 3 0 5 0
0 4 8 12
9 4 0 5
Stock Price
10 5 0 5
2 0 3 5
3 0 2 5 4
4 0 1 5
5 0 0 5 2
6 1 0 5
7 2 0 5
0
8 3 0 5 0 4 8 12
9 4 0 5 Stock Price
10 5 0 5
0 0
y ST y ST
y y
0 ST 0 ST
5 1 -1 -1 5 3
6 1 0 -1 5 2
1
7 1 1 -1 5
0
8 1 2 -1 5 -1
9 1 3 -1 5 -2
10 1 4 -1 5 0 2 4 6 8 10 12
Stock price
EduPristine FRM-I (2016) 409
Put Call Parity
0 5 0 5
1 4 1 5
2 3 2 5
3 2 3 5
4 1 4 5
12
10
Total payoff
8
6
4
2
0
0 2 4 6 8 10 12
Share price
According to Put Call parity for European options, purchasing a put option on ABC stock will be
equivalent to
A.
B.
C. ^
D.
B: p + S0 = c + Ke-rT
Put Call parity provides an equivalence relationship between the Put and Call options of a
common underlying and carrying the same strike price.
It can be expressed as:
Value of call Present value of strike price = value of put share price
If value of put is not available, it can be derived as:
Value of put = Value of call present value of strike price - share price
Put-call parity relationship, assumes that the options are not exercised before expiration day, i.e. it
follows European options.
This holds true for American options only if they are not exercised early.
In case of dividend-paying stocks, either the amount of dividend paid should be known in advance
or it is assumed that the strike price factors the future dividend payment.
The mathematical representation of Put Call Parity is:
X d
Pr emium(C ) PV of strike price PV of dividends
1 rt 1 rt
= Initial stock price (S) Put premium (P)
Put Call Parity is valid only for European options, for American Options this relationship turns into an inequality
Consider a 1-year European call option with a strike price of $27.50 that is currently valued at
$4.10 on a $25 stock. The 1-year risk-free rate is 6%.What is the value of the corresponding put
option?
A. 4.1
B. 5
C. 6
D. 25
p + S0 = c + D + y-rt
4
Call payoff
0
0 1 2 3 4 5 6 7 8 9 10 11
Share price
4
Call payoff
0
0 1 2 3 4 5 6 7 8 9 10 11
Share price
Solution
C.
The time to expiration
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
FMP-III
Options Valuation
EduPristine www.edupristine.com
Agenda
Complications in Valuing Options
Binomial Method of Valuing Options
An Example
Replicating Call Option
Replicating Put Option
Risk Neutral Valuation
Change in Future Stock Price
Generalizing Binomial Method
Black Scholes Model
Historical Volatility and Implied Volatility
Limitations of Black Scholes Model
Summary
An Example
Solution
The pay-offs are as follows:
The pay-offs are exactly the same as in the previous example for the call option. It follows that
the value of the call today should be equal to the value of 0.5714 shares less Present Value of
$36.43
Thus, value of Call = $12.86
Scenario 1 Scenario 2
Stock Price 63.75 113.33
Scenario 1 Scenario 2
Stock Price 63.75 113.33
Solution
Develop a 2 step tree.
D
Variance = 9% (120 92)2 92)2 92)2 = 168
Thus, standard deviation = 12.96
85
69.36 104.21
3 Months
-18.6%
56.6 85 127.76
6 Months
-18.6% -18.6%
If the time intervals could be made extremely small, we would be able to account for a large
number of changes in the share price
With the help of computer programs available today the binomial method can be used with very
small time intervals
If one can think of infinitesimally small time intervals, one could have a continuum of stock prices
as reflected in the plot below:
0.1 Lognormal Distribution
0.09
0.08
0.07
Probability
0.06
0.05
0.04
0.03
0.02
0.01
0
0 10 20 30 40 50 60
Stock Price
Summary
Summary (cont.)
Replicating a call option
Construct a package containing
Buy delta stocks and
Borrow a sum of money which is equal to the difference between the pay-offs from the option and pay offs
from the delta shares
This package has the same pay-off as a call option
The value of the package is the value of the call option
Summary (cont.)
Black Scholes Model
Assumes log normal distribution of stock prices
Provides a model for valuing European options on non dividend paying stocks:
R f uT
[ N (d1) u P ] [ N (d 2) u X u e ]
Where, P
ln[ ] [ R f (0.5 u V 2 )] u T
d1 X
V T
d 2 d1 V T
Question
Company X owns a property with a book value of 80,000. There is a buyer willing to pay
200,000 for the property. However, Company X must also provide the buyer with a put option to
sell the property back to Company X for 200,000 at the end of 2years. Moreover, Company X
agrees to pay the buyer 40,000 for a call option to repurchase the property for 200,000 at the
end of 2 years. In effect, with this transaction Company X borrows money from the buyer. What
is the annually compounded interest rate per year on this implied loan
A. 11.80%
B. 25.00%
C. 41.40%
D. Cannot be determined
FMP-III
Options Trading Strategies
EduPristine www.edupristine.com
Agenda
Trading Strategies
Covered Call
Protective Put
Spread Strategies
Combination Strategies
Trading Strategies
Traders may create positions using different kinds of strategies depending on the:
Expectations regarding the movement of the price of the underlying
Risk appetite
Availability of contracts
Covered Call
Involves selling call options of stocks already owned or simultaneously bought
Motivation
Earning a return from the underlying that is already owned
Lowering the cost of acquisition of the underlying asset
Expectation`
Moderate rise in the price of the underlying
Profit Potential
Maximum Profits when the options are exercised by the buyer
W^W Spot Price
If the options are not exercised the trader gets to keep the premium, thus lowering the cost of acquiring
the asset
More conservative than buying the stock only
$32 0 (100)
$33 100 0
Protective Put
Spread Strategies
0
95 100 105 110 115 120 125
-2
-4
Share Price
Bear Spread
4
Total Profit
0
95 100 105 110 115 120 125
-2
-4
Share Price
Butterfly Spread
4
Total Profit
0
100 105 110 115 120 125 130
-2
-4
Share Price
Short Call
Profit
Long Call
Diagonal Spreads
Created when both expiration date and strike price of the calls are different for Bull, Bear and
Calendar spreads. Can increase profit maximization possibilities.
Combination Strategies
Straddle
Straddle Profits
2.5
2.0
1.5
1.0
Total Profits
0.5
0
-0.5 0 2 4 6 8 10
-1.0
-1.5
-2.0
-2.5
Share Price
Strap
Strap Profits
5.0
4.0
3.0
2.0
Total Profits
1.0
0
-1.0 0 2 4 6 8 10
-2.0
-3.0
-4.0
-5.0
Share Price
Strip
Strip Profits
5.0
4.0
3.0
2.0
Total Profits
1.0
0
-1.0 0 2 4 6 8 10
-2.0
-3.0
-4.0
-5.0
Share Price
Long Strangle
Strangle Profits
2.5
2.0
1.5
1.0
0.5
0
0 2 4 6 8 10
-0.5
-1.0
-1.5
Rf Box Spread
Stock Price
Question
The S&P March 2005 index futures contract is trading at 280. The associated American 260 call
option is at 16 and the associated 260 American put option is at 3. Which of the following
strategies would you select to lock in a profit?
A. No strategy would result in a risk-free profit.
B. Buy the put, sell the call and buy the futures contract.
C. Buy and exercise the put and buy the futures contract.
D. Buy and exercise the call and sell the futures contract.
Question
An investor sells a June 2008 call of ABC Limited with a strike price of USD 45 for USD 3 and buys a
June 2008 call of ABC Limited with a strike price of USD 40 for USD 5. What is the name of this
strategy and the maximum profit and loss the investor could incur?
A. Bear Spread, Maximum Loss USD 2, Maximum Profit USD 3.
B. Bull Spread, Maximum Loss Unlimited, Maximum Profit USD 3.
C. Bear Spread, Maximum Loss USD 2, Maximum Profit Unlimited.
D. Bull Spread, Maximum Loss USD 2, Maximum Profit USD 3.
Exotic Options
Exotic Options
GAP Options: Occurs when there is a gap between the stated strike price and the strike price on
which pay-off will take place.
Eg: You buy a put option from an insurance company on your asset with a strike price of $10.0 million.
However you will incur $1.0 million cost in asset transfer, if you exercise the option. So, effectively, you would
exercise the option at <$9.0 million and not at <$10.0 million.
Forward Start Option: Options which will start at some time in the future
Eg: Employee Stock Options (ESOPs)
Cliquet Options: Consists of a series of call/ put options, where the first option is initially active
and subsequent option becomes active upon expiry of previous option.
Eg: First Option with K strike price and T time to expiry. At T first option is settled and next option becomes
active with K1 strike price (may be mutually decided at T) and T1 time to maturity.
Compound Options: Option on another option as underlying
Eg: A call option on another call option, put option on another put, a call on another put and a put on
another call.
Chooser Options: Option holder can choose whether option is call or put after a specified period
of time.
Exotic Options
Exchange Options: Option to exchange one asset for another.
Eg: Option to give up an Asset worth P at time T1 and receive in return another asset worth Q.
Volatility/ Variance swaps: Exchange of cash flows over pre-specified time intervals based on
actual realized volatility/ variance and pre-specified volatility/ variance over a notional principal.
Total Profit
Total Profit
4 4 4
2 2 2
0 0 0
95 100 105 110 115 120 125 95 100 105 110 115 120 125 100 105 110 115 120 125 130
-2 -2 -2
-4 -4 -4
Share Price Share Price Share Price
Total Profits
Total Profits
Total Profits
Thank z
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
FMP-IV
Swaps, Commodities, Foreign Exchange, Central
Counterparties, Corporate Bonds and Mortgage
Backed Securities
EduPristine www.edupristine.com
Agenda
Introduction
The Comparative Advantage Argument
Interest rate swaps
Valuation of swaps
Currency swaps
Credit risk
Other types of swaps
Introduction
A swap is an agreement between two parties to swap cash flows in the future
The arrangement covers swaps on multiple dates
Futures or forwards can be considered as a simple example of a swap where there is a cash flow
exchange on one particular date
Most common swaps are Interest Rate Swaps (IRS) and currency swaps
The legal agreement in which the two parties enter is called a confirmation, which covers the
termination date, calendar used, rates of payment, day count conventions etc.
Table tells us that X can borrow fixed at 5% and Y can borrow fixed at 7%
Also X can borrow floating at LIBOR and Y can borrow floating at (LIBOR 100bps)
This implies that X has absolute advantage in borrowing over Y
The point to note here is that the difference in fixed borrowing rates for X and Y (is not the same
for the floating borrowing rates)
Combined benefit to both X and Y by using swap is (' Fixed ' Floating ) which is 100 bps for X and Y
5.50% 5.50%
5%
X IB Y
Libor Libor >
100bps
10.00% 10.50%
10%
yz IB ABC
>/KZ >/KZ >
250bps 100bps 100bps
150bps
yz/
Floating Rate
Party 1 Party 2
Fixed Rate
Changing a liability
8%
10% >/KZ0.3%
Party 1 Party 2
LIBOR
Party 1 avails a loan of 10% while party 2 avails a floating rate loan at LIBOR 0.3%
Party 1 is receiving a fixed rate of 8% from party 2
Party 1 pays floating interest rate
Party 1s effective cash flow:
Net cash outflow is (LIBOR 2% )
Party 2s effective cash flow:
Net cash outflow is 8.3 %
Point to note here is that Party 1s fixed liability is changed to floating liability after the swap.
Party 2s liability is changed from a floating liability to a fixed liability of 8.3%
In the practical world most swaps are traded in the OTC market where financial institutions act as
market makers
LIBOR LIBOR
10% Financial
Party 1 Party 2
Institution >/KZ
9.985% 10.015%
0.3%
In the diagram above you can see that the financial institution is making a 3 basis point spread on
the fixed payment of the transaction
In such cases the bank has separate contract with party 1 and party 2
Party 1 and party 2 might not even know that they are on the other sides of the same swap
Bank creates a market by creating both bid and offer positions so that it can seek clients on either
side of the swap
It is exposed to certain credit risks in case it is unable to find a counter party for a swap
Swap rates
Valuation of swaps
There are 2 ways to value a swap.
Considering it as a difference of two bonds
Considering it as a portfolio of FRAs
Value using bonds
Consider an example in which the swap lasts for n years. If the payments are made at the end of each year
then :
If the principal is exchanged between the 2 parties at the end of the swap, then Party 1s cash flow suggests
that its long a fixed rate bond and short a floating rate bond.
Party 2 is short a fixed bond and long a floating rate bond.
We can value the swap by looking at the pay offs of either party.
Hence the value of the swap can be given as:
V = Bfix Bfl
Where:
Bfix = PV of payments
Bfl W/-rt
Value of a floating bond is equal to the par value at coupon reset dates and equals to the Present Value of
Par values (P) and Accrued Interest (AI)
Currency Swaps
Questions
Which of the following statements is correct when comparing an Interest rate Swap with a
Currency Swap?
A. At maturity there is no exchange of principal between the counterparties in IRS and there is an exchange of
principal in Currency Swaps.
B. At maturity there is no exchange of principal between the counterparties in Currency Swaps and there is an
exchange of principal in IRS.
C. The counterparty in an IRS needs to consider fluctuation in exchange rates, while currency swap
counterparties are only exposed to fluctuations in interest rates.
D. Currency swaps counterparties are exposed to less counterparty credit risk due to offsetting effect of
currency risk and interest rate risk embedded within the transaction.
Question
The USD interest rate is 4% per annum and the AUD rate is 6% per annum. Assume that the term
structure of interest rates is flat in the US and Australia. Assume current value of AUD to be $0.91.
Company ABC, under the terms of a swap agreement, pays 7% per annum in AUD and receives 3%
per annum in US$. The principal in the US is 10million USD and that in Australia is 11million AUD.
Payments are exchanged each year and the swap will last for 3 more years. Determine the value of
swap assuming continuous compounding in all interest rates.
Time Cash Flow ($) Present Value Cash Flow (AUD) Present Value
1 0.3 0.2885 0.77 0.7264
2 0.3 0.2774 0.77 0.6853
3 0.3 0.2667 0.77 0.6465
3 10.0 8.8900 11 9.2358
Total 9.7225 Total 11.2940
Credit Risk
Question
Which of the following achievable swap positions could be used to transform a floating-rate asset
into a fixed-rate asset?
A. Receive the floating-rate leg and receive the fixed-rate leg of a plain vanilla interest-rate swap
B. Pay the fixed-rate leg and receive the floating-rate leg of a plain vanilla interest-rate swap
C. Pay the floating-rate leg and pay the fixed-rate leg of a plain vanilla interest-rate swap
D. Pay the floating-rate leg and receive the fixed-rate leg of a plain vanilla interest-rate swap
FMP-IV
Swaps, Commodities, Foreign Exchange, Central
Counterparties, Corporate Bonds and Mortgage
Backed Securities
EduPristine www.edupristine.com
Agenda
Introduction
Commodity futures and forwards
The commodity Lease rate
Storage Costs and Forward Prices
Pricing with convenience
Hedging oil costs
Hedging production costs
Strip and stack hedges
V s2( t ) f ( t )
Hedge effectiveness = 1
V s2( t )
Commodity Forwards
Commodity forward prices can be described using the same formula as used for financial
forward prices
F0,T S0 e(r G )T
For financial assets, G is the dividend yield
For commodities, G is the commodity lease rate
The lease rate is the return that makes an investor willing to buy and lend a commodity
Some commodities (metals) have an active leasing market
Lease rates can typically only be estimated by observing forward prices
F0,T S0 e(r G )T
Since r is always positive, assets with G =0 display upward sloping (contango) futures term structure
A commodity loan
If you loan a commodity, you are giving up S0 today, and will get back St
If loan is fairly priced, its NPV = 0
NPV = E0(St)e-d S0
Where is required return on the commodity
Now, suppose commodity price grows at rate g, E0(ST)= S0egT
Then, NPV = S0e(g-d S
If NPV<0
this is common for commodities (supply with near-perfect elasticity)
Therefore, to make loan feasible, you would require lender to pay you the -g difference
This would get NPV back to zero
If 1 unit is loaned, you receive a lease payment of e-g)T units, and NPV = 0
Gl Dg
F0,T S0 e(r G )T
t
1
Gl r In (F0,T / S)
T
(1 r)
Gl 1
(F0,T / S)1/T
F0,T S0 e(r G )T
And if, G = c O
Solution
&
Now suppose the holder of the asset realizes a convenience yield of 2%. What is the price?
&- 0.02) 0.5 = $2.34
The futures price dropped because the cost of carrying corn dropped
FMP-IV
Swaps, Commodities, Foreign Exchange, Central
Counterparties, Corporate Bonds and Mortgage
Backed Securities
EduPristine www.edupristine.com
Agenda
Introduction
Sources of profits and losses
Unhedged foreign assets and liabilities
Balance sheet hedging
Interest rate parity
An institutions actual exposure to any given currency is its net exposure to the currency.
For example, a banks net Great Britain Pound (GBP) exposure is given by:
net GBP exposure = (GBP assets GBP liabilities) (GBP bought GBP sold)
i.e., net GBP exposure = net GBP assets net GBP bought
A positive net exposure (net long) is subject to the risk that the foreign currency will fall while a
Negative net exposure (net short) is subject to the risk that the foreign currency will rise.
Thus if the institution fails to maintain a position where net assets matches net liabilities in the
foreign currency, its exposed to the risks due to fluctuations in that currency
In foreign exchange markets, an extra dimension foreign exchange rate comes into play
increasing the volatility of net returns of the bank if unhedged
Balance sheet hedging is achieved when the financial institution matches maturity and currency in
the foreign asset-liability book
Off balance sheet hedging is done by taking a position in the forward market
Assets Liabilities
USD 10 million 7% US loans, maturity 1 year USD 20 million 5% CDs maturity 1 year
USD 10 million equivalent 12% Euro loans,
1-year maturity
What is the return for the bank if:
1. Exchange rate is unchanged
2. If exchange rate of euro has fallen from 1.2 dollars/euro to 1.05 dollars/euro
3. If exchange rate of euro has risen from 1.2 dollars/euro to 1.35 dollars/euro
In the previous example, matching the maturity duration but not the currency composition made
the returns very unpredictable. One way to minimize this risk is through balance sheet hedging.
Consider the modified balance sheet of the previous example after hedging
Assets Liabilities
USD 10 million 7% US loans, maturity 1 year USD 10 million 5% CDs, 1 year maturity
USD 10 million equivalent 12% Euro loans, USD 10 million 9% euro CDs, 1 year maturity
1-year maturity
Now even if the euro falls from 1.2 USD/euro to 1.05 USD/euro, the bank can lock in a
positive return
Steps:
1. The bank borrows USD 10 million equivalent of Euros for a year at 9%. i.e., 10/1.2 = 8,333,333 Euros
2. Pays back the Euro CD holders at the end of maturity i.e., 8,333,333*1.09=9,083,332 Euros. Euro depreciated
to 1.05 USD i.e., 9083,332*1.05 = 9,537,499 dollars. i.e., cost of funds = -4.6%
3. As calculated in previous problem, it receives an USD equivalent of 9,800,000 from the 12% euro loans
granted. i.e., -2%
Average return on assets = (-
Average cost of funds = (-
Net return = 2.3%
In the previous examples, the Euro loans have better return than US loans and lead to arbitrage
argument
As more banks move to euro loans, the spot exchange rate for buying euro will rise because of
excess demand of Euro
In equilibrium, the forward exchange rate falls to completely eliminate the attractiveness of Euro
investments. This is called Interest Rate Parity (IRP)
T
1 rDC
Forward Spot
1 rFC
A japanese investor can invest in Japanese Yen at 4.5% or in GBP at 4.67%. Current spot rate is
0.01 JPY/ GBP. Calculate 1 year forward rate in JPY/GBP.
Solution
EduPristine www.edupristine.com
Agenda
CCP is a financial intermediary which helps in trade clearing process. Clearing is a process that occurs
after execution of a trade in which a CCP may step in between counterparties to guarantee
performance.
Main function of CCP: To interpose itself directly or indirectly between counterparties to assume
their rights and obligations by acting as buyer to every seller and vice versa. This means that the
original counterparty to a trade no longer represents a direct risk, as the CCP to all intents and
purposes becomes the new counterparty.
The general role and mechanics of a CCP are:
Sets standard for its clearing members
Takes responsibility for closing out all the positions of a defaulting clearing members
Maintains financial resources to cover losses in the event of a clearing member default in the form
of:
1. Variation Margin: To closely track market movements
2. Initial Margin: To cover the worst case liquidation or close out costs above the variation margin
3. Default Fund: To mutualise (risk sharing) losses in the event of a severe default
2. Margining: Protects against the adverse movements of the market and helps uphold the settlement of the
trade.
3. Loss mutualisation: Risk sharing mechanism helps protect the financial systems from contagion of default
risk originating from a loss of one or more counterparty.
4. Liquidity: Orderly close out of the position by way of proper auctioning mechanism along with netting
reduces price impact of unwinding of a large loss making position and enhances liquidity.
Characteristics of bilateral OTC derivatives trading, their role in recent financial crisis and regulatory
changes implemented after the financial crisis:
OTC derivatives are private contracts and often contains exotic features. The underlying asset on the
derivative can be interest rate, currency etc. OTC derivatives have counterparty credit risk that is
typically pronounced during the time of default of the counterparty as unwinding of positions put
downward pressure on the already falling prices. The same may lead increased volatility and illiquidity.
This counterparty risk has been historically managed by bilateral clearing process.
The risk with bilateral clearing process is the default of a counterparty can lead to significant systematic
risk which was evident in the financial crisis of 2007 and failure of AIG which was eventually bailed out
by the government.
Clearing: Clearing is the process of reconciling and matching contracts between counterparties from
the time the commitments are made until settlement.
Forms of Clearing:
1. Direct Clearing: Means bilateral reconciliation of commitments between the original two counterparties. For
example, if party A is in contract with party B where it is required to make a payment of USD 5000 and in an
another transaction between the same party, party B is require to make a payment of USD 4500 to party A.
Under the direct clearing mechanism, rather than exchanging two cash flows of USD 5000 and USD 4500
each, party A will make a payment of USD 500 as a final settlement for both the transaction.
2. Clearing Ring: Under such mechanism of clearing, counterparty exposure is reduced between three or more
parties. It is a voluntary mechanism but once the member joins it, they have to follow the rules of the
exchange and must accept each others contracts. For example, here D owns USD 100 to B and is expected
to receive the same amount from party C. By using clearing ring, party D can be removed and party c and B
can be made counterparty. Also, party A remains unaffected in the process. Clearing ring can mitigate
counterparty risk and simplify the dependencies of a members open position and allow them to close out
contracts more easily whereby increasing liquidity.
Complete Clearing: Refers to clearing through CCP. All exchange traded contracts are currently
subject to central clearing. The CCP function can be operated either by exchange or provided to
the exchange by a third party.
CCP assumes all contractual responsibilities as counterparty to all contracts: Here party A owes
USD 125 to Party C and B owes USD 50 to party A. When CCP is introduced, party A will be
required to pay USD 75 to CCP being a central counterparty. It is evident from the below figure
that CCP clearly reduces the counterparty risk and facilitates settlement.
Comparison between exchange traded and OTC derivatives: Exchange traded derivatives are
standardized contracts with greater liquidity and regulation as compared to OTC contracts. OTC are
bilateral customized contracts which increases the hedging utility of the product by reducing basis
risk and risk of term mismatch between what is to be hedged and hedging instrument. Clearing of
OTC derivative is challenging due to its long term nature of contracts and late settlement.
Exchange traded contracts are settled through CCP reducing counterparty risk whereby OTC
contracts are generally settled bilaterally.
Exchange-traded Over-the-counter (OTC)
Terms of contract Standardised (maturity, size, Flexible and negotiable
strike, etc.)
Maturity Standard maturities, typically Negotiable and non-standard
at most a few months Often many years
Liquidity Very good Limited and sometimes very poor
for non-standard or complex
products
Credit risk Guaranteed by CCP Bilateral
It is worth noting that only notional value should not be considered to determine dominating
product as many contracts like Fixed for Floating interest rate swap never requires exchange of
principle amount which is considered notional. Hence, gross market value is a more useful indicator
to understand the derivatives market product mix. Hence, a ratio of Gross Market value to Gross
Notional Value is useful and the ratio should be relatively small and close to 3% for interest rate,
foreign exchange and credit default swap.
1. Minimizing Market risk: In terms of market risk, DPCs can attempt to be close to market- neutral
via trading offsetting contracts. Ideally, they would be on both sides of every trade as these
mirror trades lead to an overall matched book.
3. Credit risk management and operational guidelines: Restrictions are also imposed on
counterparty credit quality and activities (position limits, margin, etc.). The management of
counterparty risk is achieved by having daily mark-to-market and margin posting.
DPCs used defined triggers for their own failure through a pre packaged bankruptcy process,
which outlines the bankruptcy process and is simpler alternative to standard bankruptcy process.
Once is it is into bankruptcy, it either continues as a part of another firm or is terminated.
The question on their AAA rated status, their link with their parents (whose credit rating is worse
than itself), and advent of alternative AAA rated entities followed by global financial crisis
reduced the importance of DPCs.
Monolines are well-capitalized entities with their AAA ratings supported by capitalization
requirement based on possible losses and related to the assets for which they provide guarantee.
They are generally highly leveraged and do not have to post margin. Many monoline companies fell
during financial crisis of 2007.
CCPs generally set margin levels solely on the risks of the transactions held in each members
portfolio. Initial margin does not depend significantly on the credit quality of the institution
posting it: the most creditworthy institution may need to post just as much initial margin as others
more likely to default. Two members clearing the same portfolio may have the same margin
requirements even if their total balance sheet risks are quite different.
Auctions: CCP absorbs the Domino effect of a counterparty default by acting as a central shock
absorber and swiftly terminates all financial relations with the defaulting counterparty without
suffering any losses. CCP guarantees the performance of the trade of the surviving members by
replacement of the defaulted counterparty with one of the other clearing members for each trade.
This is typically achieved via the CCP auctioning the defaulted members positions amongst the
other members.
1. Regional
2. Product
Due to a very significant systematic role played by CCPs, they need to be resilient . Hence, a utility
led long term stable business model is preferred over short term profit maximizing goal. However, it
can also be argued that it could also be argued CCPs will need to have the best personnel and
systems to be able to develop the advanced risk management and operational capabilities.
Moreover, competition between CCPs will benefit users and provide choice. Expertise and
competition implies that CCPs should be profit-making organisations.
Disadvantages of CCPs
Moral hazard
Adverse Selection
Bifurcation: The requirement to clear standard products may create unfortunate bifurcations
between cleared and non-cleared trades. This can result in highly volatile cash-flows for
customers, and mismatches (of margin requirements) for even hedged positions.
Procyclicality: CCPs may create procyclicality effects by increasing margins (or haircuts) in volatile
markets or crisis periods. The greater frequency and liquidity of margin requirements under a CCP
(compared with less uniform and more flexible margin practices in bilateral OTC markets) could
also aggravate procyclicality.
2. Failed Auctions
3. Resignations
4. Reputational
1. Fraud
2. Operational
3. Legal
4. Investment
Liquidity Risk
Other risk:
2. Forex Risk
3. Custody Risk
4. Concentration Risk
5. Sovereign Risk
6. Wrong-way Risk
Being a central clearer, CCP creates huge concentration of risk. CCPs should not become overly
competitive during buoyant market and increase the likelihood of falling during volatile market and
crashes.
CCPs are systematically important and their failure lead to cross border impact due to global nature
of derivatives markets. The failure of CCP can be termed a bigger failure than a bank.
If the failure of a single large participant in the OTC derivatives market is capable of endangering the
entire financial system then so is the failure of a CCP that clears OTC derivatives. Therefore,
governments are likely to have little choice as to whether or not to support a failing OTC CCP. This
realisation is problematic since taxpayers bailing out a CCP is no better than bailing out other
financial institutions such as banks. Indeed, a CCP bailout represents a bank bailout of sorts since it
protects the banks that are CCP members (that may be viewed as having taken an excessively large
exposure to the CCP.
FMP-IV
Swaps, Commodities, Foreign Exchange, Corporate
Bonds and Mortgage Backed Securities
EduPristine www.edupristine.com
Agenda
Introduction to corporate bonds
Interest payment classification
Retiring of bonds before maturity
Credit risk
Default rates
Corporate Bonds
A corporate bond is a debt instrument that obligates the issuer to pay an indicated percentage of
the bonds face value on designated dates and repay the bonds face value at maturity
In the event that either the interest or principal payments are not paid, the bond is in default
Bondholders have a higher priority for the issuing companys income over preferred and common
shareholders
In the United States, corporate bonds are issued in denominations of $1,000 and multiples thereof
Indenture is a contract that states the promises of the corporate bond issuer and the rights of the
bond holder
As the indenture is hard to interpret, a third party called the Trustee is introduced
The basic functions of a trustee are:
To authenticate the bonds issued; the trustees keep a record of all the bonds sold and ensure they do not
exceed the principal amount stated in the indenture
The trustees ensure that the issuing firm adheres to all the covenants of the bonds indenture
Credit Risk
Credit Default Risk : The uncertainty whether the issuer can pay interest and principal payments
in a timely manner
Credit Spread Risk: A credit spread is the difference between a bonds yield and the yield of a
comparable maturity benchmark of a Treasury security
Credit spread risk is the risk of a loss in the value of a bond from changes in the level of credit spreads used in
the marking to market
Factors affecting credit spread risk:
Macroeconomic factors
Level and slope of treasury yield curve
Business cycle
Consumer confidence
Issue-Specific factors
Corporations financial position
Future prospects of the firm and its industry
A measure of credit spread risk is spread duration. Spread duration is the change in the value of a bond for
1% change in credit spread, assuming the underlying treasury securitys yield is constant
FMP-IV
Swaps, Commodities, Foreign Exchange, Corporate
Bonds and Mortgage Backed Securities
EduPristine www.edupristine.com
Agenda
Residential Mortgage Products and Types
Payment/Pre-payment rates
Prepayment
Prepayment Modeling
Mortgages are polled and packages to investor in Secondary Market through Securitization
Lien Status: Lien status of impacts the lenders ability to recover the balance owed in the event of
default.
First Lien is senior than other subsequent i.e. In event of default lender of first lien loan will have first
right to receive proceeds.
Original Loan Term: Loan term commonly varies from 10-30 years with 30 year loan being most
common. Medium term loan of 10-20 year maturity recently gaining popularity.
Prime (A-grade) loans: FICO score 660 or greater, Low loan-to-value ratios (<95%).
Subprime (B-grade) loans: FICO score less than 660, High loan-to-value ratios (>95%)
Alternative-A loans: Between Prime and Subprime, essentially prime loans but certain characteristics
makes then riskier (incomplete documentations, LTV on higher end)
Adjustable-rate mortgages (ARM): Variable interest payments, mostly linked to market interest rate
(LIBOR, OTC etc)
Credit Guarantees:
Every monthly mortgage payment or EMI is due on the first of each month, and consists of interest
on the outstanding mortgage balance and repayment of portion of outstanding mortgage loan B(n)
y
d B(n) u
12
y
The principal component of the payment is the remainder: EMI - B(n) u
12
Mortgage Banks
Borrower Mortgage Backed
Security
Mortgage Pool of Mortgages Banks (Investment/
Borrower (collecting principal Commercial), Mutual
repayments and Funds, Pension House,
Mortgage
interest payments, Institutional/ HNI investors,
Borrower
also prepayments) Insurance Firms, etc
Mortgage
Borrower
Securitization can be structured in numerous ways to suit the Investors requirement
Mortgage pools can be segregated into Interest-Only (IO) tranche and Principal-Only (PO) tranche
Mortgage pools can have various classes to adjust prepayments (lowest class to receive highest
prepayment, etc)
Mortgage 1 Investor 1
Mortgage 2 Investor 2
Pool
Mortgage N Investor N
e.g. calculate the prepayment for a month, if a remaining mortgage balance is $250 mn with a
next month SMM of 0.55% and the scheduled principal payment for that month is $5 mn
10
0
30
0
9
18
27
36
45
54
63
72
81
90
99
108
117
126
135
144
153
162
171
180
CPR 6% u 22 30
4.4% 150 PSA 1.5 u 4.4% 6.6%
1 1 6.6%
1
SMM 12 0.005674
Prepayment ..Cont
$1,250.00
$1,200.00
$1,150.00
$1,100.00
Prepayment Mortgage
$1,050.00
Security
$1,000.00
$950.00
$900.00
$850.00 Standard Bond
$800.00
$750.00
7.0% 8.0% 9.0% 10.0% 11.0% 12.0%
Contraction risk refers to consequences resulting from a decline in interest rates for a pass-
through security. It results in faster prepayments leading to shortening of life
Limited upside potential for a pass-through security
Reinvestment risk that received cash flow have to be invested at a much lower interest rate
Extension risk refers to consequences resulting from a increase in interest rates.
It results in slower prepayments leading to lengthening of average life
Unlimited downside potential with a limited upside makes pass-through security unattractive
Factors Causing Dollar role to trade special: Decrease in back month price (Due to access sale by
originator) and Increase in front month prices (shortage of securities due to increase in demand of
certain maturity)
Prepayment Modeling
There are four major components of prepayment modeling:
Refinancing: Using proceeds of new mortgage to pay of f principal for existing mortgage.
Factors impacting refinancing :
Interest rate fall. Also known as Media effect as large declines in rate likely gain media attention.
Cash-out Refinancing: Increase in property value allowing borrowers to get more cash for new mortgage on same
property.
Refinancing burnout: In cyclic change in interest rate, most of the refinancing occurs on first dip with less people
opting to refinancing on subsequent dips.
Incentive function: Modeling any dollar gain that borrower will refinance.
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
EduPristine www.edupristine.com
VaR Methods
Introduction to Risk
Risk can be broadly defined as the degree of uncertainty about future net returns
Credit risk relates to the potential loss due to the inability of a counterpart to meet its obligation
Operational risk takes into account the errors that can be made in instructing payments or settling
transactions
Liquidity risk is caused by an unexpected large and stressful negative cash flow over a short period
Market risk estimates the uncertainty of future earnings, due to the changes in market conditions
Broadly the standard deviation of the variable measures the degree of risk inherent in the variable
Say the standard deviation of returns from the assets owned by you is 50% and the standard
deviation of returns from assets I own is 0%. We can say that risk of my assets is zero
Say the 95% daily VAR of your assets is $120, then it means that out of those 100 days there would
be 95 days when your daily loss would be less than $120. This implies that during 5 days you may
lose more than $120 daily.
Visualizing VAR
Confidence (x%) Zy
0.45 Probability
90% 1.28
0.4
95% 1.65
0.35
97.5% 1.96
0.3
99% 2.32
0.25
95% daily-VAR
0.2
0.15
0.1
0.05
0
-4 -3 -2 -1 0 1 2 3 4
Mean = 0
The colored area of the normal curve constitutes 5% of the total area under the curve.
There is 5% probability that the losses will lie in the colored area i.e. more than the VAR number.
0.45
0.4
0.35
0.3
0.25
0.2
VARX % (in %) Z X % *V
0.15
0.1
0.05
0
-4 -2 0 2 4
Mean = 0
y
0 and standard deviation as 1)
VAR in absolute terms is given as the product of VAR in % and Asset Value:
VAR VARX % (in %) * Asset Value
This can also be written as:
VAR Z X % * V * Asset Value
This comes from the known fact that the n-period volatility equals 1-period volatility multiplied by
the square root of number of periods(n).
As the volatility of the portfolio can be calculated from the following expression:
V portfolio wa2 V a2 w 2b V b2 2w a w b * V a * V b * U ab
The above written expression can also be extended to the calculation of VAR:
Solution
Daily VaR = 0.016 x 10 x 2.33 = 0.3728 mn
Solution
Solution
VaR(A)(in %) = 5.5 x 1.96 = 10.78%; VaR(B)(in %) = 4.25 x 1.96 = 8.33%;
Question 4
Question 5
VaR
Linear: When the value of the delta is constant for any change in the underlying
Primarily in the case of forwards and futures we have linear assets
The method to calculate VAR for linear assets is called Delta Normal method
Delta Normal method assumes that the variables are normally distributed
Non Linear: When the value of the delta keeps on changing with the change in the underlying
asset
Options are non-linear assets, where delta-normal method cannot be used as they assume the linear payoff
of the assets
To calculate the VAR for non-linear assets, full revaluation of the portfolio needs to be done
Monte Carlo methods or Historical Simulation are commonly used to fully revaluate the portfolio
linear
Payoff
0
k Share/asset price
Long position
Delta of Derivative: Change in price of Derivative to change in underlying asset
For example:
The permitted lot size of S&P CNX Nifty futures contracts is 200 and multiples thereof.
So VAR of Nifty Futures contract is 200 * VAR of Nifty Index
If the daily VaR at 5%of Nikkei is USD 0.8 mn and you have 100 lots of Nikkei contract.
Calculate annual VaR at 95% confidence for your portfolio assuming 250 days?
Here, delta = 100, because for every 1 unit change in the Index Nikkei, the futures price will change
by 100 units because the lot size is 100
The fat-tailed unconditional distribution can be broken down into two conditional distributions,
either with similar means and different variances or similar variances and different means.
Many a times when we observe marked differences between the estimated and actual volatilities,
its a result of regime switching which means that the average volatility in the market has now
changed too much when compared to the previous estimate
Regime Switching
1 n
V2 (R i E ( R)) 2
n i1 Actual volatility
30bp/day
18bp/day
Low Vol Estimate
Unconditional
1 n
volatility V2 (R i E ( R)) 2
n i1
10bp/day
x
n
2
i x
i 1
Variance
n 1
Mean of returns (x-bar) is usually zero, especially if returns are over short-time period (say, daily returns). In
that case, variance estimate for next day is nothing but simple average (equally weighted average) of
previous n days squared returns.
n
x
2
i
i 1
Variance
n 1
What if the volatility is dependent on the values of volatility observed in the recent past?
What if they also depend on the latest returns?
Hence we have:
m i 1 2
V 2
n (1 O ) O u n i OmV n2 m
i 1
Variance estimate for next day (n) is given by (1-
previous variance estimate.
Risk-metrics (by JP Morgan) assumes a Lambda of 0.94.
In this equation, variance for time t was also an estimate. So we can substitute for it as follows:
V t21 (1 0.94) Pt2 0.94>(1 0.94) Pt21 0.94V t21 @
Example 1: On Tuesday, return on a stock was 4%. Volatility (Std. deviation) estimate for Tuesday
was 1%. Find volatility estimate for Wednesday using of 0.94
EWMA Solution
Variance estimate for Wednesday = (1-
Std. Dev. = sqrt (1.9%) = 1.378%
Tuesday volatility (Std. Dev.) estimate was 1%. Actual return on Tuesday was 4%. Therefore,
volatility estimate for Wednesday is estimated upwards than Tuesday i.e. 1.378% as compared to
1%.
Notice how the volatility estimate has been revised due to high return
Example 2: Continuing the previous example, volatility estimate for Wednesday was 1.378%.
Assume that actual return on Wednesday was 0%. What is the variance estimate for Thursday?
EWMA Solution
,^&
hZDtD
conditional variance, which weight will be applied to the return that is 4 days old?
A. 0.000
B. 0.043
C. 0.048
D. 0.950
Solution
B.
A. Incorrect. The wrong factor has been squared. The EWMA RiskMetrics model is defined as:
ht t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of (1-
0.95)3*0.95 = 0.00012 for r0 when t = 4.
B. Correct. The EWMA RiskMetrics model is defined as:
ht t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of (1-
0.95)*0.95^3 = 0.043 for r0 when t = 4.
C. Incorrect. The 0.95 has not been squared. The EWMA RiskMetrics model is defined as:
ht t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of (1-
0.95)*0.95 = 0.048 for r0 when t = 4.
/ddtDZD
ht t-1 - 2t-1. For t = 4, and processing r0 through the equation three times produces a factor of 0.95 =
0.950 for r0 when t = 4.
V t21 JV L DP t2 EV t2
Long-term average Volatility
V t21 Z DP t2 EV t2
's>
Since the sum of all the weights is equal to 1 we get the following equation as well:
Z
VL
1 D E
GARCH Solution
Solution: In the GARCH model, 12% is the weight given to latest squared return (reactive factor).
85% is the weight given to latest variance estimate (persistence factor). Therefore,
1-0.12-0.85 = 3% is weight given to long-term average Volatility.
Therefore, 3%*VL = 0.000005 i.e. VL = 0.017%
Also, variance estimate for 1 = .000005 0.12*(-1%)^2 0.85*(1.88%)^2 = 0.0317%
Volatility (Std. Dev.) estimate for 1 = sqrt (0.0317%) = 1.782%
For a stable GARCH model, alpha Beta <=1. If alpha Beta>1, then weight given to long-term volatility is
negative and the model becomes mean-fleeing
Solution
B.
A. Incorrect. The model that will take the shortest time to revert to its mean is the model with the lowest
1 /
1 0.03 0.96 = 0.99.
B. Correct. The model that will take the shortest time to revert to its mean is the model with the lowest
1 /
1 0.02 0.95 = 0.97.
C. Incorrect. The model that will take the shortest time to revert to its mean is the model with the lowest
1 /
1 0.01 0.97 = 0.98.
D. Incorrect. The model that will take the shortest time to revert to its mean is the model with the lowest
1 /
1 0.01 0.98 = 0.99.
Solution
The new volatility is 1.53% per day
Ans -2733814418
HS Ans -2.35
Portfolio Volatility
Third Approach
Aggregate the simulated returns and then apply the parametric approach to aggregated portfolio
Truly Predictive in nature and is not a historical simulation but it assumes that Black Scholes
(or any other model) is correct model for calculating market Value of Option
Solution
99% of 500 = 495th i.e. -5.9% = 0.059 x 70 = 4.13
Solution
VaR = 396th = -6.1% of 100 = -6.1
Expected shortfall = - -6.7%
Full Valuation method is the process of measurement of risk of a portfolio by fully re-pricing it
under a set of scenarios over a time period. It can be used to cover a large range of values of the
portfolio returns in order to provide more accurate results. It generally provides more accurate
results compared to delta normal approach but is a complicated process.
Two popular methods under full revaluation approach have been explained in the subsequent
slides.
Payoff
Gains !
Price
Risk !
Disadvantages:
The calculation of Monte Carlo VaR can take 1,000 times longer than Parametric VaR because the potential
price of the portfolio has to be calculated thousands of times.
Unlike Historical VaR, it typically requires the assumption that the risk factors have a Normal distribution.
Visualizing WCS
The worst case scenario (WCS) answers:
What is the worst loss that can happen over a period of time?
The probability of a worst loss is certain (100%); the timing is uncertain.
Focuses on distribution of loss during the worse trading period:
A worst case scenario can be simulated only using Monet Carlo Simulations
The area under the normal curve for confidence value is:
Confidence (x%) Zy
90% 1.28
95% 1.65
97.5% 1.96
0.005 Mean = 0
- stdev
Question (WCS)
H = 100
-2.33 ( 99% VaR ) 1 out of 100
Distribution of WCS mean = -2.51 and its 1st and 5th percentile are -3.1 and -3.9 resp.
What does this mean to a financial risk manager?
Stress Testing
Limitations of VaR
Tells that n number of times in 100 days, the loss is not going to exceed N$
But it cannot predict the loss when it exceeds!
Does not focus on large losses (Tails of distribution)
Stress Testing:
Supplement to VaR
"VaR should always be supplemented with stress testing" has been one of the recommendations
of the supervisor
Testing how well a portfolio performs under some of the most extreme market moves seen in the last 10 to
20 years
Stress Testing
Advantages:
Can take a large number of risk factors into consideration
Can specifically focus on the tails (extreme losses)
Disadvantages:
Highly subjective and can become overcautious
Requires complete top management support
Capital Allocation
Exposure
Ride out Turmoil
The one significant shortcoming of VaR that stress testing does address is sudden changes in
historical correlations
If two currencies have been pegged to one another, they will exhibit a high historical correlation. A
VaR analysis based on that historical correlation will not address the risk that one of the currencies
may be devalued relative to the other. If this is a scenario that concerns management, a simple
stress test will offer more insights than would, say, a VaR analysis performed with a modified
correlation assumption
Scenario Analysis
Goals
Identify scenarios that would not occur under standard VAR models
Simulating shocks that have never occurred
Simulating shocks that reflect permanent structural breaks or temporally changed statistical patterns
Factor push method shifts each risk factor in the direction that would have an adverse impact on
the portfolio.
Conditional Scenario Method incorporates the correlation between various key risk factors.
Stress tests can be improved by adopting one or more of the following methods:
Buy protection through insurance or other derivative products
Change portfolio composition to decrease the exposure or diversify
Change business strategy to suit the changing business environment
Develop back up plans for unforeseen events
Secure alternative funding in stressed environment
Implied Volatility V 2
n
(R
i 1
i E ( R )) 2
V n2 OV n2 1 (1 O )u n21 90% 1.28
Regime Switching 95% 1.65
The worst case scenario
GARCH (1,1): 97.5% 1.96
Stress Testing
VAR Z X % * V * Asset Value
V 2
t 1 JV L DP EVt
2
t
2
99% 2.32
VaR (n days) (in %) VaR (daily VaR) (in %) * n VARLinear Derivative ' *VARUnderlying Risk Factor
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
EduPristine www.edupristine.com
Greeks
Agenda
Stoploss Strategy
Trader can hope to construct a stop-loss strategy by combining the two strategies
Suppose the trader is able to ensure that he has a covered call position each time the price of the underlying
exceeds the strike price on the option and
A naked call position whenever the stock price goes down from the exercise price. If a trader sells call
contracts standalone, he has a naked position exposure
If this was possible in the real world, traders would be able to earn riskless profits as the cost of
setting the hedge would always be less than the theoretical price of the option
However, it is impossible to execute this in reality because of the following reasons:
Purchase of stock will always be slightly above the strike price and sale slightly below it
This difference would increase the cost of this stop-loss strategy.
In case one wants to decrease this difference to a really small amount, frequent buying and selling will be
required, which in turn would again increase the cost of this trade
The difference in time of the associated cash flows and the time value of money also result in increased costs
Delta
Options delta is defined as the rate of change in an option's price relative to a one-unit change in
the price of the underlying asset
The delta for European Call option is given by N(d1) and for the Put option it is given by N(d1)-1
Holding delta share and selling one call lead to a risk neutral outcome, other things remaining
same
This is an example of delta hedging
Such a package of option and share is called a delta neutral portfolio
Delta hedging strategy is an improvement over the simple stop-loss strategy
However, the options delta changes with changes in stock price
Requires traders to frequently readjust their positions in order to remain delta neutral. This process is called
rebalancing
Delta hedging therefore is a dynamic hedging strategy
0.8 2.0
0.6 1.5
0.4 1.0
0.2 0.5
0 0
0.2 0.4 0.6 0.8 1 1.2 1471013161922252831343740434649
-0.2 -0.5
-0.4 -1.0
-0.6 -1.5
Delta (cont.)
The delta of a portfolio of derivatives (such as options) with the same underlying asset, can be
found out if the deltas of each of these derivatives are known
n
' portfolio W ' i i
Example: i 1
Lets find the delta of the following portfolio constructed from the options of the same underlying
asset and also estimate the trade required in order to turn the portfolio delta neutral
The delta of the portfolio is given by:
50,000 * 0.53 100,000 * 0.47 25,000 * (-0.51) = -7750
It follows from the calculations above that 7,750 stocks need to be bought in order to get a delta
neutral portfolio
Delta (cont.)
Theta
The theta of an option is the rate of change in its value with the passage of time, assuming that
other things remain the same
For a portfolio, the theta is the rate of change in the value of the portfolio as time passes, given
that other things are constant. A positive theta implies that the portfolio value will increase as
the time passes, while a negative theta implies that the value will decrease with the passage of
time, if there is little move in the stock price and the implied volatility
Normally expect the theta of an option to be negative as with the passage of time, an option
loses value
Exception can be an in-the money European put option on a non dividend paying stock
For at-the-money option, theta increases as the expiration date nears
Theta decreases as an option which is either out of money or in the money approaches expiration
Theta (Call)
Theta (Put)
Theta (cont.)
e ( x ^ 2) / 2 ^
N ' ( x)
2S K = Strike price
Gamma
Gamma is the rate of change in delta for unit change in the price of the stock
For a portfolio of options with the same underlying asset, the gamma is the change in the value of
the portfolios delta with the change in the portfolio value:
A large gamma suggests that delta will change rapidly as the price of the underlying stock changes
The following table presents and compares the sign of Gamma and Delta
Calculation of Gamma
Gamma for European options can be calculated using the following formula:
N ' ( d1)
*
S 0V T
Where symbols have their usual meaning
Gamma (cont.)
Vega
The Vega of a derivative portfolio is the rate of change of the value of the portfolio with the
change in the volatility of the underlying assets. It can be expressed as:
G3
s
GV
For European options on a stock that does not pay dividends, Vega can be found by:
s^dE^dE
given by:
e ( d 1^ 2 ) / 2
N ' ( d1) Vega
2S 16
The Vega of a long position is always positive 14
12
A position in the underlying asset has a zero Vega
10
Thus its behavior is similar to gamma 8
Vega is maximum for options that are at the money 6
4
2
0
1 4 7 1013161922252831343740434649
Rho
Rho of a portfolio of options is the rate of change of its value with respect to changes in the
interest rate
Rho = G 3 , where is the value of the portfolio, and r is the rate of interest
Gr
For European options on non dividend paying stocks, we have;
Rho (call) = KTe-rTN(d2), where the symbols carry their usual meanings
Also, Rho (put) = -KTe-rTN(-d2), the symbols carrying their usual meanings
Summary
Traders may combine naked and covered positions to evolve a stop-loss strategy
Stop loss strategy is impractical given the realities of trade and transaction costs involved
Delta hedging is an improvement
Involves that delta of the portfolio is maintained at zero
Requires frequent rebalancing as delta change
Dynamic hedging strategy
Hedging can also be attempted with respect to changes in:
Time (theta)
Option delta itself (Gamma)
Volatility (Vega)
Interest rate (Rho)
In addition to option Greeks, traders also rely on scenario analysis
Involves evaluating the option value for simultaneous changes in:
Time (Theta)
Volatility (Vega)
Interest rate (Rho)
Other factors
1. A stock trading at $20 has call options available on it with exercise prices $18 and $20. For $1
increase in the stock price how will the delta of the two options change? Choose the most
appropriate answer. Change in deltas for the two options are denoted by d '18 and d '20.
A. d'18 < d'20
B. d'18 > d'20
C. d'18 = d'20
D. d'18 > d'20 and d '20 = 0
Question
1. A.
2. C.
Vega is the rate of change in the price of an option with respect to changes in the volatility of the underlying
asset. Vega is greatest for at-the-money options with long times remaining to expiration
3. D.
Theta is large and negative for an atthe-money European-styled option, whilst theta is close to zero when
the price for the underlying stock is very low. Therefore the theta for an out-of themoney European styled
call option would have a lower negative value compared to that of an at-the-money European-styled call
option
Operational Risk
Some companies define operational risk as all the risk that is not market or credit risk. But this is a
very broad definition of operational risk
The Basel definition of operational risk is the risk of direct or indirect loss resulting from
inadequate or failed internal processes, people and systems or from external events
Operational Risk definition includes legal risk but does not include reputational or strategic risk
because they can be difficult to quantify
There are three main approaches for calculating operational risk capital requirement:
The basic indicator approach
In this approach, operational risk capital is based on 15% of the banks annual gross income over the period
of 3 years
Includes both interest and non interest income
The standardized approach
In this approach, bank uses eight business lines with different beta factors to calculate the capital
requirement
Beta factor for each business line is multiplied with the annual gross income of the business line over the
period of 3 years
The results are then added to arrive at total operational risk capital charge
The advanced measurement approach (AMA)
Large banks are encouraged to move from standardized approach to AMA because with AMA, banks can
reduce their capital requirement by investing in risk management infrastructure
Expected Unexpected
Loss (EL) Loss (UL)
Probability
of loss
Potential Loss
The Basel committee on Banking Supervision divides operational risk into seven types:
1. Clients, products and business practices
2. Internal Fraud
3. External Fraud
4. Damage to physical assets
5. Execution, delivery and process management
6. Business Disruption and System failures
7. Employment Practices and Workplace safety
Loss frequency and loss distribution are combined to simulate the expected loss distribution.
This is known as convolution.
Monte Carlo Simulation process is used to run this simulation.
Data Limitations
The historical operational risk loss data is currently inadequate to run these simulations.
Banks should use both internal and external data to estimate the severity of losses.
In scenario analysis, different scenarios are created to incorporate the events that have not
yet occurred.
Scenario analysis act as a tool for the management to immunize against potential operational risk.
The drawback of scenario analysis is the excessive amount of time spent by the management to
create such scenarios.
In addition to scenario analysis, forward looking approaches are also used to discover potential
operational risk events. There are three forward looking methods:
1. Causal Relationships
2. Risk and Control Assessment
3. Key Risk Indicators
Manager of each business unit allocates operational risk capital to manage operational risk.
Less capital will be allocated to those business units which are able to reduce the frequency and
severity of losses.
One of the approaches to allocate capital is scorecard approach.
Each manager is surveyed regarding the key features of each type of risk.
Answers of the managers are given a score to quantify the responses.
Power law is very useful in evaluating the nature of the tails of a given distribution in extreme
value theory.
This law is very useful in analysis of operational risk because most of the operational risk losses are
likely to occur in the tails.
Insurance
There are many insurance companies which provide insurance for operational risk.
There are two main issues faced by such insurance companies:
Adverse Selection
Moral Hazard
Issuer Credit Rating Information about the entity that has issued the instruments.
Credit Risk Moody's* Standard & Poor's* Fitch IBCA** Duff & Phelps**
Investment Grade
Highest quality Aaa AAA AAA AAA Triple A = Credit risk almost zero
High quality (very strong) Aa AA AA AA Safe investment, low risk of failure
Most speculative Ca CC CC CC
No interest being paid or
C C C C
bankruptcy petition filed Bankruptcy or lasting inability to make
In default C D D D payments most likely
* The ratings from Aa to Ca by Moody's may be modified by the addition of a 1, 2 or 3 to show relative standing within the category.
**The ratings from AA to CC by Standard & Poor's, Fitch IBCA and Duff & Phelps may be modified by the addition of a plus or minus sign to show relative.
1. Qualitative Analysis
2. Quantitative Analysis
3. Meeting the firms management
4. Committee in the credit rating agency meets to assign the rating
5. Assigned rating is notified to the firm
6. Opportunity for the firm to appeal or offer new information
7. Disseminating the rating to the public via the news media
One-year ratings migration probabilities based upon bond rating data from 1981-2000. Data is
adjusted for rating withdrawals. Numbers in each row should sum to 100%. Due to round-off
error, they may not do so exactly. Source: Standard & Poor's
The chances of default given any rating at the beginning of a cycle increases with the time horizon.
External ratings are designed to be relatively stable over the business cycle. There is a possibility
that there are errors in severe cycles.
But external ratings have fairly good records in indicating relative defaults rates.
At-The-Point (ATP) Assesses for one year e.g. scoring models, KMV etc.
Through-The-Cycle (TTC) Captures the credit-worthiness over a larger period of time, including
the impact of normal cycles in the economy, these ratings are more stable.
Some banks use TTC for large corporate and ATP for SME.
One view is that TTC is the combination of various ATP at various time.
Issues in Back-Testing
There is no database when an internal rating system is designed.
Generally a time of 10 to 18 years required to validate an internal rating system.
APT will have pro-cyclicality if they stick to PD but will have unstable transition matrices and no long term
guidance if they follow internal ratings.
Agenda
What is country risk?
Sources of Country risk
Measuring Country risk
Sovereign defaults
Consequences of sovereign default
Factors affecting sovereign default risk
Sovereign Rating
Criticism of ratings agencies
Sovereign Default Spreads
The need to understand, analyze and incorporate country risk has become a priority at most large
corporations, as they have globalized and become more dependent upon growth in foreign markets for
their success. Thus, a chemical company based in the United States now has to decide whether the hurdle
rate that it uses for a new investment should be different for a new plant that it is considering building in
India, as opposed to the United States, and if so, how best to estimate these country-specific hurdle rates.
Governments often have a direct effect on country risk, with increased country risk often translating into
less foreign investment in the country, leading to lower economic growth and potentially political turmoil,
which feeds back into more country risk.
3. Physical violence
4. Nationalization/Expropriation risk
Legal risk
Economic Structure: Well diversified economy or reliance of economy on single commodity
Measurement model/methods
No Standardization
More ranking than scores: The country risk score are more useful for ranking country on risk rather than
measuring risk. A risk score of 60, compared with a risk score of 30 is definitely more risky but we can not
conclude that it is double risky.
Sovereign Default
Risk that government issued bonds will fail in their contractual repayment of coupon payments or maturity
value or both in timely manner to the investors.
1. Countries have been more likely to default on bank debt owed than on sovereign bonds issued.
2. In dollar value terms, Latin American countries have accounted for much of sovereign defaulted debt in
the last 50 years
Gold Standard: Till 1971, countries following the gold standard, had to be backed up with gold
reserves. As a consequence, the extent of these reserves put a limit on how much currency could
be printed.
Shared Currency: When the Euro was adopted as the common currency for the Euro zone, the
countries involved accepted a trade off. In return for a common market and the convenience of a
common currency, they gave up the power to control how much of the currency they could print.
Thus, in July 2015, the Greek government cannot print more Euros to pay off outstanding debt.
Inflationary concerns
Sovereign Ratings
Bond rating agencies have initiated sovereign rating and they have two big advantage in gaining acceptance
in the market:
1. Rating agencies have been assessing default risk in corporations for a hundred years or more and
presumably can transfer some of their skills to assessing sovereign risk.
2. It is easy for investors to relate the assessment of risk conveyed by the rating agencies as they are
familiar with the ratings measures, from investing in corporate bonds. The same leads to a AAA rated
country to be viewed as close to riskless whereas a C rated country is very risky.
Rating agencies generally provide both local currency and foreign currency rating and generally
local currency ratings are at least as high or higher than foreign currency rating. The same is due to
more power in the hands of the government to print its own currency and pay-off the debt.
Sovereign ratings do change over time but less frequently than corporate rating. The same is
evident from the rating transition matrix.
Fitch sovereign Transition rates across the major rating categories: 1995-2008 (%, Annual)
The rating agencies have been criticized d for having a regional bias means they under rate entire regions
of the world (Latin America and Africa). However, the argument of rating agency for the same is past
default history is a good predictor of future default and the regions rated lower have a bad history to
overcome default.
Rating measure: Sovereign ratings focus on credit risk to private creditors and not to official creditors like
IMF, World Bank etc. Default means either a failure to pay interest or principal on a debt instrument on
the due date (outright default) or a rescheduling, exchange or other restructuring of the debt
(restructuring default).
Determinants of Rating
Political Risk:
Stability and legitimacy of political institutes
Participation of population in political process
Leadership Succession
Transparency in economic policy decisions and objectives
Public security
Geopolitical risk
Economic Risk:
Market orientation of economy
Income disparities
Financial intermediation effectiveness: Availability of credit
Labour flexibility
Rating Process: The analyst with primary responsibility for the sovereign rating prepares a ratings
recommendation with a draft report, which is then assessed by a ratings committee composed of 5-
10 analysts, who debate each analytical category and vote on a score. Following closing arguments,
the ratings are decided by a vote of the committee.
Local versus Foreign currency rating: The ratings agencies usually assign two ratings for each
sovereign a local currency rating and a foreign currency rating. There are two approaches used by
ratings agencies to differentiate between these ratings. In the first, called the notch-up approach,
the foreign currency rating is viewed as the primary measure of sovereign credit risk and the local
currency rating is notched up, based upon domestic debt market factors. In the notch down
approach, it is the local currency rating that is the anchor, with the foreign currency rating notched
down, reflecting foreign exchange constraints.
The differential between foreign and local currency ratings is primarily a function of monetary policy
independence. Countries that maintain floating rate exchange regimes and fund borrowing from
deep domestic markets will have the largest differences between local and foreign currency ratings,
whereas countries that have given up monetary policy independence, either through dollarization or
joining a monetary union, will see local currency ratings converge on foreign currency ratings.
Rating review and updates: Sovereign ratings are reviewed and updated by the ratings agencies and
these reviews can be both at regular periods and also triggered by news items. Thus, news of a
political coup or an economic disaster can lead to a ratings review not just for the country in
question but for surrounding countries (that may face a contagion effect).
When a government issues bonds, denominated in a foreign currency, the interest rate on the bond can be
compared to a rate on a riskless investment in that currency to get a market measure of the default spread
for that country. To illustrate, the Brazilian government had a 10-year dollar denominated bond
outstanding in July 2015, with a market interest rate of 4.5%. At the same time, the 10-year US treasury
bond rate was 2.47%. If we assume that the US treasury is default free, the difference between the two
rates can be attributed (2.03%) can be viewed as the markets assessment of the default spread for Brazil.
Even though there is strong co relationship between sovereign ratings and market default spreads, there
are advantages for using default spread for assessment of default risk. The primary advantages are:
The first is that the market differentiation for risk is more granular than the ratings agencies; thus, Peru
and Brazil have the same Moodys rating (Baa2) but the market sees more default risk in Brazil than in
Peru.
The second is that the market-based spreads are more dynamic than ratings, with changes occurring in
real time.
Exercise
Q.1 An analyst makes following comment with regard to the measurement of country risk. Identify
which one of the following is correct.
2. Country risk score of 80 in comparison of a score of 40 can be concluded as double risky country.
3. Political Risk Service (PRS) uses three dimensions to assess country risk.
4. Under the economists country scoring system, low score means higher risk and high score means low
risk.
A. BBB is countrys local currency rating and BBB- is its foreign currency rating.
B. Local currency rating is generally higher or equal to foreign currency rating due to flexibility of
country to print its own currency.
3. Either of A or B is correct
Exercise
Q.3 Sovereign default will not lead to
2. Rating downgrade
4. Political change
A. Rating agencies may walk in tandem with action of other rating agency compromising
independence.
1. A, B and C
2. A and C only
3. C only
4. A and B only
Answers
A.1: Option 3
A.2: Option 1
A.3: Option 3
A.4: Option 4
Credit Risk
Credit losses leads to economic losses, consisting Expected Loss and Unexpected Loss
EL = (Assured payment at Maturity Time T )* Loss Given Default * (Probability that the default
occurs before maturity Time T)
The term Assured payment is more relevant for bonds than loans
For Bank Loans the terms Assured Payment is better replaced by Exposure
EL = Exposure * LGD*PD
EL is the amount the bank can lose on an average over a period of time
However, during the same period because of macroeconomics conditions, or market trends there
could be some deviation from the expected loss which is known as Unexpected Loss (UL). Banks
need capital to take care of this loss
Unexpected Loss
UL is the estimated volatility of the potential loss in value of the asset around its EL
UL is the standard deviation of the unconditional value of the asset at the time horizon
UL = s.d. of expected asset value
h>&>'>'&
Underlying assumption that EDF is independent of LGD. In case it is not so then correlation between LGD
and EDF terms will come into picture. Though it has been found that they will affect the result only slightly
What if all the variance were zero?
Thank z
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
Foundation of Risk Management - I
EduPristine www.edupristine.com
Agenda
Risk Management: A Helicopter View
Corporate Risk Management: A Primer
Corporate Governance and Risk Management
What is ERM?
Risk-Taking and Risk Management by Banks
Financial Disasters
The Credit Crisis of 2007
Risk Management Failures: What are they and When do they Happen?
What is Risk?
Risk, in traditional terms, is viewed as a negative. The dictionary defines risk as exposing to danger
or hazard.
You cannot have opportunity without risk and that offers that look too good to be true (offering
opportunity with little or no risk) are generally not true. By emphasizing the upside potential as well
as the downside dangers, this definition also serves the useful purpose of reminding us of an
important truth about risk.
Assess costs
t & benefits of
Assess effects of Exposures
Instruments
Evaluate Performance
Source: Risk Management: A Helicopter View
Unexpected Loss
A catastrophic loss of a magnitude well beyond losses seen in the course of normal daily life. The real risk is
that the daily life costs will suddenly rise in an entirely unexpected way, or some other cost will appear from
nowhere and steal the money that was kept aside for our annual budget.
Portfolio
Operational Risk Issue Risk
Concentration
Credit Risk
Legal & Regulatory
Transaction Risk Issuer Risk
Risk
Risks
Reputation Risk
William Sharpe establishes that firms should not worry about the risks that are specific to them
because all the specific risks are diversified away in a large investment portfolio, under the perfect
capital markets assumption.
With the effect of progressive tax rates, volatile earnings induce higher taxation than stable
earnings.
Risk management activities allow management better control over the firms natural economic
performance.
Risk reduction activity offers synergies with the operations of the firm (Hedging the price of an
input of a firm can stabilize its cost and pricing policy, which in turn offer a competitive advantage
in the marketplace that could not be replicated by any outside investors)
If a company chooses to hedge activities related to its operations, such as hedging the cost of raw
materials, this clearly has implications for its ability to compete in the marketplace.
For example When an American manufacturing company buys components from French company in euros,
the American company can opt to avoid the foreign currency risk by hedging the exposure.
This hedge has both a size and a price effect-i.e., it might affect both the price and the amounts of products
sold.
Two conclusions
Firms should risk-manage their operations
Firms may also hedge their assets and liabilities, so long as they disclose their hedging policy
At most banks, the board charges its main committees with ratifying the key policies and associated
procedures of the banks risk management activities. The various committees are:
Responsible for the accuracy of the banks financial & regulatory
reporting and also ensure that the bank complies with best-
Audit Committee
practice standards in key activities such as regulatory, legal,
compliance and risk management
Risk Committee of the Board Approves market risk tolerance each year
What is ERM?
The Committee of Sponsoring organization of the Treadway Commission (COSO) in 2004 defined
ERM as
ERM is a process, effected by an entitys board of directors, management, and other personnel,
applied in strategy setting and across the enterprise, designed to identify potential events that may
affect the entity, and manage risk to be within its appetite, to provide reasonable assurance
regarding the achievement of entity objectives.
Risk is the effect of uncertainty on objectives and risk management refers to coordinate activities
to direct and control an organization with regard to risk.
Benefits of ERM
Three major benefits of ERM are:
Increased
Improved business
organizational Better risk reporting
performance
effectiveness
Establishing an integrated risk management framework for all aspects of risks across the
organization
Developing risk management policies, including the quantification of the firms risk appetite
through specific risk limits
Implementing a set of risk indicators and reports, including losses and incidents, key risk
exposures, and early warning indicators
Allocating economic capital to business activities based on risk, and optimizing the companys risk
portfolio through business activities and risk transfer strategies
Developing the analytical, systems, and data management capabilities to support the risk
management program
Corporate Governance
Establish top-down risk management
Line Management
Business strategy alignment
Portfolio Management
Think and act like a fund manager
Risk Transfer
Transfer out concentration or inefficient risks
Risk Analytics
Develop advanced analytical tools
Stakeholders
Management Improve risk transparency for key stakeholders
ERM is an integral part of organizations culture and is embedded in its corporate strategy,
planning and day to day decision making
As part of ERM, an organization must consider its risk appetite, alongside its goals and operational
tactics. Management should take three steps w.r.t. risk appetite:
1. Develop risk appetite
2. Communicate risk appetite
3. Monitor and update risk appetite
As per COSOs Enterprise Risk Management Integrated Framework definition, Risk Appetite:
Is the amount of risk, on a broad level, an entity is willing to accept in pursuit of value. It reflects
he entitys risk management philosophy, and in turn influences the entitys culture and operating
style.
Is strategic and is related to the pursuit of organizational objectives
Forms an integral part of corporate governance
Guides resource allocation.
Assists the organization in aligning the organization, people and processes in designing the
infrastructure necessary to effectively respond to and monitor risks.
Influences the organizations attitude towards risk
Is multidimensional
Requires effective monitoring of the risk itself and of the organizations continuing risk appetite
Risk Appetite
Considerations affecting risk appetite
1. Existing Risk Profile current level and distribution of risks across the entity and across various
risk categories
2. Risk Capacity The amount of risk the entity is able to support in pursuit of its objectives
3. Risk Tolerance Acceptable level of variation an entity is willing to accept regarding the pursuit of
its objectives
4. Attitudes towards Risk The attitudes towards growth, risk and return
1. Develop Management develops, with broad review and concurrence, a view of the
organizations overall risk appetite
2. Communicate The view of risk appetite is translated into a written or oral form that can be
shared across the organization
3. Monitor Management monitors the risk appetite over time adjusting how it is expressed as
business and operational conditions warrant
Good Risk: Provides ex-ante reward on standalone basis i.e. Value creating risk opportunities
Bad Risk: Leads to loss of value without providing an opportunity for value creation. This type of risk
are danger and needs to be avoided.
While assuming good risk, bank will increase its overall risk as bank will invest in assets and projects
that are valuable but can also lead to a loss in value because of an adverse impact which will
increase its risk of financial distress and its ability to create value through liabilities. Hence, a well
governed bank will have to identify an optimal level of risk from its shareholder perspective. Bank
should try not to deviate much from its optimal level of risk amount.
To make it simple, bank should take up the project that increases its value while taking into
consideration the cost associated with the project on total risk of the bank. However, practically, risk
taking decisions are taken throughout various departments of the bank almost continuously and
hence, each risky decision can not be considered in isolation but must be assessed in terms of its
impact on overall risk of the bank.
So, should banks always target a AAA rating? The answer is No. Because, by targeting the highest
credit rating bank will also limit its opportunity in taking up projects which can increase shareholder
wealth. A target of optimal credit rating for a bank is a function of composition of its business
activities.
For example, A bank with more of a deposit franchise and with more relationship lending is likely to
prefer a higher rating than an institution that is engaged in more transactional activities. Similarly, a
bank that enters in long-term derivatives contracts might find a higher rating more valuable than
one that does not because the counter partys assessment of credit risk is paramount in such type of
long term contracts.
Hence, the ratings that maximizes the value of the bank differs across banks.
2. Bank Safes value falls steeply if the bank is riskier than its target rating and increases only
moderately as it increases its risk towards the target rating. Bank Risky has a substantially
different relation between value and its rating. Its target rating is BBB and its value increases
substantially as it increases its risk towards its target and falls sharply if it exceeds it. For both
banks, having too much risk is extremely costly in terms of their value. However, for one bank
having too little risk has little cost while for the other it has a large cost.
The risk management function of the bank should be evaluated with the value it generates in terms
of protecting the bank from taking excessive risk.
Also, flexibility of risk management function is paramount so as not to merely control the risk but
also provides framework for exploiting value creating opportunities.
Risk management function should take the perspective of the entire bank level risk rather than
individual level department risk. Individual business units are generally not capable of determining
that whether their risk taking will add or destroy value at the firm level. Hence, risk management
function can add value by requiring all the business units to take a firm wide perspective while taking
risk.
2. Hedging limitations
Due to the above limitations, risk has to be managed and monitored through out the organization.
Bank divides its risk into category of Market, Credit and Operational risks. A firm wide VaR,
aggregating all the categories of risk typically does not capture all the risks .i.e. business risk
Business risk, non-interest income of bank (processing fees etc) unexpected changes in interest rates
and unexpected changes in credit spread needs to be modelled separately.
VaR also does not capture funding liquidity risk. A shock to funding can force the bank to sell assets
at a loss. Further, shocks to funding are more likely to happen in periods when markets for securities
are themselves less liquid, so that selling assets quickly will be costly as they are sold at a discount.
Hence, the same needs to be modelled separately.
Different types of risk will lead to different statistical distributions and typically credit and
operational risk have non-normal distributions and is difficult to model them.
Risk that are unknown, black swan or unknown unknowns are important only when VaR is estimated
at extremely low probability levels like 0.06% VaR. A focus on historical data and the use of
sophisticated statistical technique is not much useful for developing forecast for this extremely rare
events. A stress testing technique can be applied to find whether bank will withstand the
occurrence of those rare events.
1. There is limited data available about how the risk function is organized in firms.
2. Banks risk appetite has also greater influence on banks risk function apart from governance
3. At a firm level, poor ex-post performance can be consistent with very good risk management
The concept of risk capital needs to be considered while designing compensation for the managers.
Risk Culture:
Research concludes that: Companies employing honest and trustworthy managers are more
profitable and were given high valuation.
Shareholder governance improvement would change a firms culture from focusing on employee
integrity and customer service to focusing on end results.
Financial Disasters
Inexperienced Managers:
Capital: $20 Mn 1. Thought they were just
Borrowed middlemen
2. Didnt realize contract
Debt Market: $300 Mn indicated Chase taking
(Unsecured Loan) full responsibility of
debt
LTCM was a hedge fund using highly leveraged arbitrage trading activities in fixed income
in addition to pairs trading. Before failing in 1998, it had given spectacular returns in 1995-97
periods (up to 40% post-fees). Post Russian default on its ruble denominated debt, LTCM lost more
than USD 4bn in 4 months
LTCM used proprietary mathematical models to engage in arbitrage trading in US, Danish, Russian,
European and Japanese Govt. bonds. In 1998, LTCMs positions were highly leveraged (1:28) with
~US$5: 130 Bn of equity and assets
LTCMs model assumed maximum volatility of 20% annually. Based on its models, it was expected
to losses more than ~US$500 Mn in once in 20 months
It had its bet on convergence of Russian and American G-sec yield, which however diverged
after Russian default. Its failure led to a huge bailout by large commercial and merchant banks
under the guidance of Federal Reserve
It had various risk exposures .such as Model Risk, Funding liquidity risk, Sovereign Risk,
Market Risk
Following the Enron episode, Investment banks and regulators agreed to put new controls in place,
to make sure that their clients were accounting derivative transactions with them, in ways
transparent to investors.
However the controls were not followed strictly. Goldman Sachs helped Greece in hiding its level
of indebtedness from its EU partners, by disguising debt as an interest rate swap.
The swap was done deliberately at an off market rate, creating an upfront payment to Greece, that
was scheduled to be paid back by Greece with suitable interest over the course of Swaps life.
However, effectively the transaction was a loan arrangement, which was disguised as a derivative
transaction.
In 1991 Kidder, Peabody & Co. hired Joseph Jett to their Strips trading desk
In his first 6 months the desk did not make money, then reported profits grew quickly ($32 Mn in
1992,
$151 Mn in 1993, and $81 Mn in the first quarter of 1994)
Jett was Kidders man of the year in 1993 and was awarded a $9 million bonus (his boss received
a $20 Mn bonus in 1993, primarily for overseeing Jetts activities)
It was later reported that Jetts desk lost $85 Mn during that period and a $358 Mn cumulative
write-down. The false profits were the result of an error in their internal accounting whereby a
zero-coupon bond was treated like a coupon bond. Thus a forward reconstitution, which
exchanges a zero for an identical coupon bond, created instant profits
Beware of increasing trading volume. Kidders balance sheet was $25 Bn in 1991, $273 Bn in 1992,
and $1,567 billion in 1993. Any time notional amount grows dramatically it is wise to double check
the business model
Most value-destroying activity is not illegal. Kidder and GE shareholders brought suit against Jett
for fraud, but it was impossible to prove because Jett was openly engaged in his trading fraud
Compensation is an imperfect signal of competence
Credentials are imperfect signals of competence. Joseph Jett had an MIT undergraduate degree
and a Harvard MBA. Jetts boss, Ed Cerullo, had a respectable career on Wall Street, and was
considered street smart and technically savvy
The back office and auditors should not be fearful of traders, and it is managements job to make
this so
1993: John Rusnak, who had been working for Chemical Bank in New York, joins First Maryland
Bancorp as a foreign exchange trader
1999: Allfirst is formed from the merger of First Maryland Bancorp (in which AIB first took a stake
in 1983) and Dauphin Deposit Corporation (which AIB acquired in 1997). Susan Keating becomes
Allfirst chief executive
June 2001: John Rusnak is promoted to managing director in charge of foreign exchange trading, in
the global trading division of the treasury funds management section, or front office
Late December, 2001: Allfirst officials start to become suspicious about the sums being demanded
by Rusnak to cover his trading
January 10, 2002: Keating is appointed to the AIB chief executive committee, the group
responsible for developing corporate strategy and overseeing management of AIB group
February 4, 2002: Rusnak fails to show up for work on Monday morning
February 6: AIB says it is investigating a suspected $750 Mn fraud at Allfirsts Baltimore HQ, and
warns that it will take a one-off charge of E596 Mn ($520 Mn) to cover the resulting losses
February 8: Eugene Ludwig, a former US Comptroller of the Currency, is hired to compile a report for AIB on
the affair
February 19: AIB chief executive Michael Buckley says that the origins of the scandal might stretch
back to 1997, and gives the final figure for losses as $691 Mn
March 12: Buckley and AIB chairman Lochlann Quinn offer their resignations to the AIB board, but
neither resignation is accepted
March 13: The Ludwig Report is published jointly by Ludwigs Promontory Financial Group and law
firm Wachtell, Lipton, Rosen & Katz
March 14: Allfirst and AIB announce that six executives who were responsible for oversight of
Rusnaks activities are to be dismissed. A number of organizational and structural changes are also
announced, including the appointment of an individual to oversee risk management across the AIB
group
March 17: AIB denies rumors that Allfirst CEO Keating is about to step down. Later newspaper
reports claim that Keating has been given one year to get Allfirst back on track
The originators of mortgages did not in many cases keep the mortgages themselves. They sold
portfolios of mortgages to companies that created products for investors from them. This process
is called securitization.
The structure of Asset-Backed Securities
Assets
Principal: $100 Company ABS
million
Senior Tranche
Principal: $75 million
Return = 6%
Equity Tranche
Principal: $5 million
Return = 30%
Assets Cash
Flows
Senior Tranche
Mezzanine
Tranche
Equity Tranche
Junior AAA 5%
AAA 81%
AA 3%
AA 11%
Subprime Mortgages
A 2%
A 4%
BBB 1%
NR 1%
Senior AAA 60%
Senior AAA 62%
Junior AAA 27%
Junior AAA 14%
AA 4%
AA 8%
BBB 3% A 3%
A 6%
BB/NR 1% BBB 3%
BBB 6%
NR 2%
NR 4%
The Crisis
In general, a firm will specify a risk measure that it focuses on together with additional risk metrics
(Example for risk metric is VaR)
When that risk measure exceeds the firms tolerance for risk, risk is reduced
Alternatively, when the risk measure is too low for the firms risk tolerance, the firm increases its
risk
The managers took risks they should not have, but that is not a risk management issue as long as
the risks were properly understood
Rather, it is an issue of assessing the costs of losses versus the gains from making large profits.
Example: Failure of LTCM
With good risk management, large losses can occur when those making the risk-taking decisions
conclude that taking large, well understood risks creates value for their organization
Risk Metrics like VaR, help measure and quantify risk and help managers to take decisions in line
with the targeted risk measures
Problems with VaR:
Frequency of Reporting: Daily losses below VaR, but accumulating on an annual basis
Does not capture extremely large losses which have a very low probability of occurrence
VaR assumes distribution of losses is not correlated. Correlation among various variables increases in times
of crisis
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com
EduPristine www.edupristine.com
Agenda
The Standard Capital Asset Pricing Model
Applying CAPM to Performance Measurement: Single Index Performance Measurement Indicators
Arbitrage Pricing Theory & Multifactor Models of Risk and Return
Information Risk and Data Quality Management
Principles for Effective Risk Data Aggregation and Risk Reporting
GARP Code of Conduct
Market risk premium: Additional return over the risk-free rate, required to compensate investors
for assuming an average amount of risk. Its size depends on the perceived risk of the stock market
and investors degree of risk aversion
Rs R f R m R f
covR i , R m
i
Var R m
Return
SML
Rf
Beta
Linear relation of risk and return plotted on a graph is called Security Market Line (SML)
SML determines - if an asset is expected to offer commensurate return w.r.t. potential risk
Single stock return is plotted versus its security's risk
If it is above SML, stock is undervalued because market expects higher return
If it is below SML, stock is overvalued because market expects lesser return
Beta
Sensitivity of the return of the asset to the market return is known as Beta
Beta is calculated as follows:-
covR i , R m
i
Var R m
Portfolio Beta
Beta can also be calculated for portfolio
Portfolio Beta is the weighted average of the betas of individual assets in the portfolio
Jensons alpha R p R c
Rp = portfolio return, Rc = return predicted by CAPM
Positive alpha (portfolio with positive excess return) is always preferred over negative alpha
Measures of performance
Tracking Error
V R p R b
Information Ratio
E ( R p ) E ( Rb )
V R p R b
Sortino Ratio
E ( R p ) MAR
1 T
Nt0
R p MAR
2
For a given portfolio, the expected return is 10% with a standard deviation of 15%. The beta of the
portfolio is 0.75. The expected return of the market is 11% with a standard deviation of 18%.
The risk-free rate is 4%. The portfolio's Treynor measure is:
A. 0.060
B. 0.012
C. 0.040
D. 0.080
D.
TR
0.10 0.04 0.08
0.75
Where Ep = RP RB
RP = portfolio return, RB = benchmark return
Lower the tracking error lesser the risk differential between portfolio and the benchmark index
Information Ratio (IR):
Measure of risk-adjusted return for a portfolio, defined as expected active return per unit of tracking error
IR
R p Rb
TE
Higher IR indicates higher active return of portfolio at a given risk level
For the past four years, the returns on a portfolio were 6%, 9%, 4%, and 12%. The corresponding
returns of the benchmark were 7%, 10%, 4%, and 10%. The minimum acceptable return is 7%.
The portfolio's Sortino ratio is:
A. 0.4743
B. 0.2143
C. 0.5303
D. 0.6700
A.
Average Return 6 9 4 12 7.75%
4
SR
0.0775 0.07
SSD
SSD
1 0 3 0
2 2
10
1.58%
4 4
0.75%
SR 0.4743
1.58%
B.
0.82
0.122 0.119
SSD
Semi Standard Deviation = SSD = 0.3658%
Semi Variance = SSD2 = 0.1338%
A portfolio has an average return over the last year of 13.2%. Its benchmark has provided an
average return over the same period of 12.3%. The portfolios standard deviation is 15.3%, its beta
is 1.15, its tracking error volatility is 6.5% and its semi-standard deviation is 9.4%. Lastly the risk
free rate is 4.5%. Calculate the portfolios Information Ratio (IR)
A. 0.569
B. 0.076
C. 0.138
D. 0.096
C.
IR
0.132 0.123 0.138
0.065
Correlation PB = 0.961
SR
> 6% 3%@0.36
25%
Absolute performance is poor
SR
> 10% 3% @ 0.65
20%
Absolute performance is even poorer than portfolio
Information ratio
Absolute performance is even poorer than portfolio
IR
> 6% 10%@ 0.50
8%
There are three recent risk adjusted performance measures which are covered in FRM
Part 1 curriculum:-
Morning Star Rating System
VaR Based Risk Adjusted Measures
Management Related Risk Adjusted Measures
Risk adjusted performance of the investment funds is ranked by the firm Morning Star
This ranking is done within a peer group
Top 10% of the funds in the peer group get 5 star rating whereas bottom 10% of the funds get 1
start rating
Middle 35% of the funds are assigned 3 star rating
VaR states the maximum loss that the asset can sustain given a stated confidence level and
time period
By modifying the Sharpe Ratio, risk-return performance based on VaR can be analyzed
R p Rf
VaR
/Vp
p
Where
VaRP = Portfolio VaR
VP = Initial Portfolio Value
VaR can also be used for investment decision making process. If the fund manager wants to add a
new security in the portfolio, then the new VaR of the portfolio will be compared with the VaR of
the portfolio before adding the security. Change in the VaR of the portfolio from the addition of
the new security is known as Incremental VaR.
In management related risk adjusted measures, capital market line is modified to capture the
appropriate investment style of the fund
This measure is also known as style risk adjusted performance measure (SRAP)
There is another management related risk adjusted measure known as correlation adjusted
performance measures
This measure has been developed to compare managers within a peer group
This measure helps the investors to allocate their funds optimally across fund managers that
maximizes the desired return and minimizes risk
What if, rather than using a single market proxy, we consider each macro factor as a separate
source of risk?
Factor Model
The systematic factor summarized by market return can arise from a number of separate factors such as
uncertainty of business cycle, interest rate, inflation rate etc.
Can be used to describe and quantify each factor that has an impact on the security return. And building
model of equilibrium security pricing
Useful in risk management. Factor model can be used to measure exposure of the stock to various risk
factors.
ri E (ri ) E i F ei
E (ri ) E i ( Factual Fexp ected ) e i
Note:
i. F is a random variable with an expected value of zero, i.e. if there are no macro surprises, (F = 0), than
ii. Ei is uncorrelated with F
Multifactor Model
ri E (ri ) E i1 F1 E i2 F2 ei
An example
ri E ( ri ) E iGDP GDP E iIR IR e i
here
E iGDP is the factor beta for unexpected changes in GDP
E iIR is the factor beta for Interest Rate
For the two factor model, the multi factor SML states, that the expected return on the security is
sum of
i. Risk free rate of return
ii. Sensitivity to GDP risk time the risk premium for bearing GDP risk
iii. Sensitivity to interest rate risk times the risk premium for bearing interest rate risk
Arbitrage
An arbitrage opportunity arises when an investor can earn riskless profit by constructing a zero
net investment portfolio. Example shares of stocks sold for different prices in different
exchanges
The law of one price states that if two assets are equivalent in all economically relevant respects,
than they should have the same market price. As per the law of one price, when arbitrage
opportunities arise, large amount of trade will be generated (buying cheap and selling expensive),
that puts pressure on security prices.
The arbitrage trades will increase the price, where it is low and decrease it where it is high, till
there is no more arbitrage opportunity available
APT vs CAPM
APT and CAPM serve similar purpose, i.e. they provide the benchmark for determining the
expected rate of returns for security valuation
APT does not have the restrictive assumption of CAPM i.e. as per CAPM, the benchmark
portfolio in the SML should only be the (unobservable) market portfolio.
In APT, any portfolio on the SML line can serve as the benchmark portfolio. For example, even if
the index portfolio is not an exact proxy of market portfolio, it cannot be used for SML (by CAPM).
But as per APT, this index portfolio can be used for SML.
Multifactor APT
Multifactor APT generalizes the single factor model to accommodate several sources of systematic
risk.
The multifactor SML predicts that exposure to each risk factor adds to securitys total risk
premium
Poor Data can impact your financials: increase operating costs, increase penalties, delays in
cash flows
Loss of Confidence of management reporting, low confidence in forecasting
Customer Satisfaction goes down
Impacts Credit Assessment
Impacts Productivity negatively
Compliance is jeopardized
Accuracy
Completeness
Consistency
Reasonableness
Currency
Uniqueness
Other Dimensions of Data Quality
Operational Data Governance is the manifestation of the processes and protocols necessary to
ensure that an acceptable level of confidence in the data effectively satisfies the organizations
business needs
It governs the roles, responsibilities associated with the quality of data
Scoreboards
A quality scoreboard:
Hierarchical rollup of metrics should reflect in the scorecards
Definitions
Appropriate level of presentation depending on the intended audience
Principles
The principles are initially addressed to SIBs and at both the banking group and on a solo basis.
Code of Conduct
1. Fundamental
1. Professional Integrity
Responsibilities
and Ethical Conduct
2. Adherence to generally
2. Conflicts of Interest
accepted practices of
3. Confidentiality
risk management
dz
help@edupristine.com
www.edupristine.com
EduPristine www.edupristine.com