CRICOS 00099F Financial Modelling and Analysis
Financial Modelling and Analysis
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
CRICOS 00099F
Financial Modelling and Analysis
Seminar 1
Introduction and Descriptive statistics
2Communication
• Content related questions
o Discussion Board in Canvas
• Administrative questions and consultation hours
o Via email to the Subject Coordinator
o [email protected]
Seminar 1 - Introduction and Descriptive statistics
3Subject contents
• Seminar 1 Descriptive statistics
• Seminar 2 Statistical inference – Confidence intervals
• Seminar 3 Statistical inference – Hypothesis Testing
• Seminar 4 Forecasting
• Seminar 5 Measuring forecasting accuracy
• Seminar 6 Linear regressions – Simple and Multiple
• Seminar 7 Linear regressions – Time series
• Seminar 8 Linear regressions – Model evaluation
• Seminar 9 Linear regressions – Additional topics
• Seminar 10 Excel modelling – Best practices
• Seminar 11 Excel modelling – Applications
• Seminar 12 Review
Seminar 1 - Introduction and Descriptive statistics
4Seminar structure - Wednesdays
In-person
2. In-class activity 3. Q&A + Quiz1. Lecture
12:00 – 12:50 13:00 – 13:50 14:00 – 14:50
Online 18:00 – 18:50 19:00 – 19:50 20:00 – 20:50
Seminar 1 - Introduction and Descriptive statistics
5Assessment
• In-class quizzes 20%
o 50 minutes with MCQs
o Quiz #1: Week 5 | Topics in weeks 1 - 4
o Quiz #2: Week 8 | Topics in weeks 5 - 7
• Case Study 20%
o Individual
o Due on Friday 12 May 2023
• Final exam 60%
o Covers the whole 11 lectures
Seminar 1 - Introduction and Descriptive statistics
6What Do I Need to Do to Succeed?
• 10 – 15 minutes of reading before coming to the seminars
• Resolve any question ASAP, e.g. during the same lecture.
• Following the in-class examples and activities
• Self-assessment: can I explain the concept or method to others?
Seminar 1 - Introduction and Descriptive statistics
Descriptive Statistics
• Population, sample, and probability
• Expectation, variance, covariance, and correlation of
random variables
• Population values when probability distribution is
known
• Sample values when probability distribution is
unknown
8Population versus Sample
• Population contains ALL potential observations.
o Parameters (Greek letters): mean μ, variance σ2, correlation ρ, etc.
• A sample contains a portion of the population.
o Estimates (Roman letters): mean ത, variance s2, correlation r, etc.
• Statistic inference: how close is ത to μ, s2 to σ2, etc.?
Population
Sample 2:
Y1,…,Yn
Sample 1:
X1,…,Xm
Seminar 1 - Introduction and Descriptive statistics
9Random Variables and Probability Distributions
• The value of a random variable is subject to chance (or randomness)
o Future values are unknown
o Degree of randomness: short-term interest rate and wage are “deterministic” compared
to asset prices and volatility.
• A probability distribution specifies all the possible outcomes along with the
probability of occurrence of each outcome
Seminar 1 - Introduction and Descriptive statistics
10
Measures of Central Tendency
Expectation
• Weighted average:
μ = E X =
i=1
M
XiP(Xi)
where
• Xi = the i
th outcome of the (discrete) random variable X
• P(Xi)= probability of occurrence of the i
th outcome of X
• M = number of possible outcomes/values
• σi=1
M () is the sum over i = 1, …, M
Seminar 1 - Introduction and Descriptive statistics
11
Measures of Dispersion
Variance and Standard Deviation
• Variance of discrete random variable X
σ2 = σi=1
M Xi − E X
2 P(Xi)
• Standard deviation of discrete random variable X
σ = σi=1
M Xi − E X 2P(Xi)
Seminar 1 - Introduction and Descriptive statistics
12
IPO Distribution Example
• Consider the following table on the number of IPOs per month on the ASX since 2010:
IPOs per month Probability
0 0.35
1 0.25
2 0.2
3 0.1
4 0.05
5 0.05
Seminar 1 - Introduction and Descriptive statistics
13
IPO Distribution Example
• The expected number of IPOs per month is
E X = σi=1
M XiP(X = Xi)
= 0 × 0.35 + 1 × 0.25 + 2× 0.20 + 3 × 0.10 + 4 × 0.05 + 5 × 0.05
= 1.40
Seminar 1 - Introduction and Descriptive statistics
14
IPO Distribution Example
• Since the mean = 1.4, the Variance is
σ2 =
i=1
M
Xi − E X
2 P(Xi)
σ2 = 0 − 1.4 2 × 0.35 + 1 − 1.4 2 × 0.25 + 2 − 1.4 2 × 0.20 +
+ 3 − 1.4 2 × 0.10 + 4 − 1.4 2 × 0.05 + 5 − 1.4 2 × 0.05
σ2 = 2.04
• Standard Deviation is
σ = σ2 = 2.04 = 1.428
Seminar 1 - Introduction and Descriptive statistics
15
Two Funds Exercise
• Consider two alternative investments. The first is an aggressive fund with high returns
during economic boom. The second is a defensive fund that performs better during
economic downturn.
• The following table summarizes your estimate of the annual returns under three
economic conditions, with each with probability occurrence:
Seminar 1 - Introduction and Descriptive statistics
16
Two Funds Exercise
Probability
Economic
Condition
Aggressive Fund
(X)
Defensive Fund
(Y)
0.2 Recession -30% 20%
0.5 Stable Economy 10% 5%
0.3
Expanding
Economy
25% -10%
Seminar 1 - Introduction and Descriptive statistics
17
Two Funds Exercise
X = Aggressive Fund
• E X = −0.3 0.2 + 0.1 0.5 + 0.25 0.3
= 0.065 = 6.5%
Y = Defensive Fund
• E Y = 0.2 0.2 + 0.05 0.5 + −0.1 0.3
= 0.035 = 3.5%
Seminar 1 - Introduction and Descriptive statistics
18
Two Funds Exercise
X = Aggressive Fund
• σX
2 = Var(X) = 0.2(−0.3−0.065)2+ 0.5(0.1−0.065)2+
0.3(0.25−0.065)2
= 0.0375
• σX = 0.194
Seminar 1 - Introduction and Descriptive statistics
Y = Defensive Fund
• σY
2 = Var(Y) = 0.2(0.2−0.035)2+ 0.5(0.05−0.035)2+
0.3(−0.10−0.035)2
= 0.011
• σY = 0.105
19
Measures of Association
Covariance and Correlation
• The covariance and correlation measure the strength of the linear relationship between
two random variables X and Y.
• A positive covariance/correlation implies that X and Y often move in the same direction.
• A negative covariance/correlation implies that X and Y often move in the opposite
direction.
Seminar 1 - Introduction and Descriptive statistics
20
Measures of Association
Covariance
• Covariance:
Cov X, Y = E{ Xi − E X Yi − E Y }
= σi=1
M Xi − E X Yi − E Y P(XiYi)
where P(XiYi) = prob. of occurrence of i
th outcome of X and ith outcome of Y.
o The sign indicates the direction of the relationship
o The size/magnitude depends on the size/magnitude of X and Y.
Seminar 1 - Introduction and Descriptive statistics
Cov(X,Y) > Cov(V,W) does not necessarily imply greater co-movement between X and Y
3.5%
Defensive Mean
Return
6.5%
Aggressive Mean
Return
21
Measures of Association
Correlation Coefficient
• Correlation coefficient:
ρXY = Cor(X, Y) =
Cov(X, Y)
σXσY
o The sign indicates the direction of the relationship
o The size/magnitude is -1 < ρXY < 1 and does not depend on the size/magnitude of X
and Y.
Seminar 1 - Introduction and Descriptive statistics
Cov(X,Y) = 25 > Cov(V,W) = 0.1
σX= σy= 8 σV= σW= 0.4
Cor(X,Y) = 0.391 < Cor(V,W) = 0.625
22
Two Funds Exercise
• Cov X, Y = −0.30 − 0.065 0.20 − 0.035 0.2
+ 0.10 − 0.065 0.05 − 0.035 0.5
+ 0.25 − 0.065 −0.10 − 0.035 0.3
= −0.0193
• ρXY =
σXY
σXσY
=
−0.019
0.194×0.105
= −0.9476
When the return of the aggressive fund is high, the return of the defensive fund is low;
and vice versa.
Seminar 1 - Introduction and Descriptive statistics
23
Measures of Association
Autocovariance
Cov(Xt, Xt+k) = E{ Xt − E Xt Xt+k − E Xt+k }
• The value depends on k, not t:
Cov(Xt−k, Xt) = Cov(Xt−k+1, Xt+1)
= Cov(Xt−k+2, Xt+2) = …= Cov Xt, Xt+k
• The sign indicates the direction of the relation
• The size/magnitude depends on the size/magnitude of X.
Seminar 1 - Introduction and Descriptive statistics
24
Measures of Association
Autocorrelation
ρX(k) = Cor Xt, Xt+k =
Cov(Xt,Xt+k)
σtσt+k
σt = var(Xt) and σt+k = var(Xt+k)
• The sign indicates the direction of the relationship
• -1 < ρX(k) < +1
Seminar 1 - Introduction and Descriptive statistics
Cov(Xt,Xt+1) = 25 > Cov(Yt,Yt+1) = 0.1
σX,t= σx,t+1= 8 σY,t= σY,t+1= 0.4
rX(1) = 0.391 < rY(1) = 0.625
Sample and Sample Statistics
26
Sample and Sample Statistics
• Probability is not directly observable
o Often we assume a population distribution
• Alternatively we can take a sample from the population and estimate sample statistics
o Central Tendency: the way data clusters around some central value
o Dispersion: the way observations tend to spread out around some central value
o Association: between two variables or two time points for the same variable
Seminar 1 - Introduction and Descriptive statistics
27
Measures of Central Tendency
Sample Mean
• The mean, or average, is the most commonly used measure of central tendency
ഥX =
1
N
σi=1
N Xi
o ഥX = sample mean
o Xi is the i
th observed value of X
o N = number of observations (sample size)
• Mean values are affected by the minimum or the maximum values in the sample
Seminar 1 - Introduction and Descriptive statistics
28
The IPO Example
• We don’t observe the population distribution of the monthly IPOs.
• We observe a 2-year sample.
• The average monthly IPOs:
o ഥX2015 = 1.33
o ഥX2016 = 1.83
o ഥX2015−16 = 1.58
Seminar 1 - Introduction and Descriptive statistics
Month
IPOs per
month
Jan-15 0
Feb-15 1
Mar-15 0
Apr-15 2
May-15 3
Jun-15 2
Jul-15 1
Aug-15 1
Sep-15 3
Oct-15 2
Nov-15 1
Dec-15 0
Jan-16 0
Feb-16 1
Mar-16 1
Apr-16 3
May-16 0
Jun-16 3
Jul-16 2
Aug-16 4
Sep-16 2
Oct-16 5
Nov-16 1
Dec-16 0
29
Sample vs Population
• In the IPO example, the population mean μ = 1.4 (slide 13) under the assumed
distribution
o Number of possible values M = 6
o Each value/outcome has an assumed probability.
o μ = 1.4 is a description of the unobserved population.
• The (full) sample mean ഥX = 1.58.
o Number of observations N = 24.
o Many observations have the same value.
o ഥX = 1.58 is a description of the sample from 2015-16.
Seminar 1 - Introduction and Descriptive statistics
30
Measures of Central Tendency
Sample Median
• The median is the middle value in an ordered array of numbers.
o Half of the observations will be smaller and half will be larger than the median.
• Median is the middle value not affected by the minimum or the maximum values.
Seminar 1 - Introduction and Descriptive statistics
31
Measures of Central Tendency
Calculation of the median
• To find the median first order the N observations from smallest to largest:
X(1) ≤ X(2) ≤ ··· ≤ X(N)
o if N is odd = +1
2
o if N is even =
1
2
2
+
2
+1
Seminar 1 - Introduction and Descriptive statistics
32
Median Example
• The price-earnings ratios of six listed companies are:
10.3 , 4.9 , 8.9 , 11.7 , 6.3 , 7.7
• The ordered sequence is:
4.9 , 6.3 , 7.7 , 8.9 , 10.3 , 11.7
• Positioning point =
6+1
2
= 3.5 (N = 6)
• The median is the average of the third and fourth ordered observations:
=
7.7 + 8.9
2
= 8.3
Seminar 1 - Introduction and Descriptive statistics
33
Median Exercise
• What is the median of the following data?
10 , 2 , 3 , 2 , 4 , 2 , 5
• Ordered array:
X(1) … X(7)
2 , 2 , 2 , 3 , 4 , 5 , 10
• Location of the median = (7+1)/2 = 4
• Therefore, the median equals 3.
• What is the median of 30 , 2 , 3 , 2 , 4 , 2 , 5 ?
o Is the mean the same?
Seminar 1 - Introduction and Descriptive statistics
34
Measures of Dispersion
Sample Standard Deviation
• The standard deviation measures the dispersion around the mean and is calculated as:
sX =
σi=1
N xi − തx 2
N − 1
o There is a greater contribution to s as x moves away from the mean
• Variance is the square of the standard deviation.
o Neither measure can be negative
Seminar 1 - Introduction and Descriptive statistics
35
Measures of Association
Sample Covariance and Correlation
• Sample covariance
Cov X, Y =
σi=1
N (Xi − ഥX)(Yi − ഥY)
N − 1
• Sample Correlation
Cor , =
Cov(X, Y)
sXsY
Seminar 1 - Introduction and Descriptive statistics
36
Linear Correlation
• The above definitions of covariance and correlation measure the strength of linear
relationship between two random variables.
o They do not measure non-linear relationships.
• For example:
o Let X = -5,…,0,..,5.
o Therefore X2 = 25,…,0,…25
o Cor(X,X2) = 0
o No linear correlation
Seminar 1 - Introduction and Descriptive statistics
37
Measures of Association
Sample Autocorrelation
• For a time series of X1,…,Xt,…,XT, the sample autocorrelation (AC) of order k is given by
=
1
T − 1
σt=1
T−k(Xt − ഥX)(Xt+k − ഥX)
1
T − 1
σt=1
T (Xt − ഥX)2
• The X’s are assumed to be taken from the same population, and have the same mean
and variance.
=
σt=1
T−k(Xt − ഥX)(Xt+k − ഥX)
σt=1
T (Xt − ഥX)2
Seminar 1 - Introduction and Descriptive statistics
38
Two Funds Sample
• In the fund return example (slide 16), if the observed time-series returns are:
Seminar 1 - Introduction and Descriptive statistics
Observation
Aggressive Fund
(X)
Defensive Fund
(Y)
1 -30% 20%
2 10% 5%
3 25% -10%
• What are the measures of central tendency, dispersion and association for this sample?
39
Two Funds Sample
Sample Mean and Standard Deviation
• Mean
ഥX =
−0.30 + 0.10 + 0.25
3
= 0.0167 = 1.67%
ഥY =
0.20 + 0. 05 − 0.10
3
= 0.05 = 5.00%
• Standard Deviation
=
−0.30 − 0.0167 2 + 0.10 − 0.0167 2 + 0.25 − 0.0167 2
3 − 1
= 28.43%
=
0.20 − 0.05 2 + 0.05 − 0.05 2 + −0.10 − 0.05 2
3 − 1
= 15.00%
Seminar 1 - Introduction and Descriptive statistics
40
Two Funds Sample
Sample Covariance and Correlation
• Sample covariance
Cov X, Y =
(−0.3 − 0.0167)(0.2 − 0.05) + (0.1 − 0.0167)(0.05 − 0.05) + (0.25 − 0.0167)(−0.1 − 0.05)
3 − 1
= −0.0413 < 0
• Sample correlation
Cor(X, Y) =
−0.0413
0.2843 × 0.1500
= −0.9672
• It is a reasonable estimate of the population correlation ρXY = -0.9476 (slide 22)
Seminar 1 - Introduction and Descriptive statistics