Professional Documents
Culture Documents
Quantitative Analysis
C FA
CF A
WWW.GFEDU.NET
400-700-9596
Framework
Framework
2-30
R5 C FA
CFA
Time Value Calculation
3-30
Interest rate
Required rate of return
Discount rate
Opportunity cost
Nominal risk-free rate=real risk-free rate + expected inflation rate
Required interest rate on a security=nominal risk-free rate + default risk
A
premium + liquidity risk premium + maturity risk premium
EAR
C F
m
r
EAR 1 1
F
m
C A
continuous compounding: EAR er 1
The greater the compounding frequency
The greater the EAR will be in comparison to the stated rate
The greater the difference between EAR and the stated rate
Annuity
Ordinary annuity
Annuity due (1) BGN(2) ordinary annuity
Perpetuity
4-30
C FA
R6
F A
Discounted Cash Flow Applications
C
5-30
NPV&IRR
N
(1 r )
CF1 CF2 CFN CFt
NPV CF0 ...
(1 r )1 (1 r )2 (1 r )N t
t 0
N
(1 IRR)
CF1 CF2 CFN CFt
NPV 0 CF0 ...
(1 IRR )1 (1 IRR )2 (1 IRR )N t 0
t
CF A
Multiple solutions Problem of the IRR calculation (#sign changes)
Independent Projects:
Accept it if NPV>0
Accept it if IRR>r (required rate of return)
Decision Rule Mutually Exclusive Projects:
NPV method : Choose the one with higher NPV
IRR method : Choose the one with higher IRR
NPV and IRR methods may conflict with each other,NPV
6-30
r ( F P0 ) 360
BD
F t
360
rMM HPY
P1 P0 CF1
t BEY 2
HPY
P0 EAY (1 HPY ) 365 / t
1 (1+
2
) =1+EAR
TWRR& MWRR
C FA
TWRR
MWRR
CF A
HPR
nHPR MWRR=IRR
MWRR
7-30
R7 C FA
F A
Statistical Concepts and Market Return
C
8-30
Types of measurement scales
C FA
arithmetic meanweighted meangeometric meanharmonic mean
CF A
Mean
harmonic mean <=geometric mean <= arithmetic mean
arithmetic mean
or ending valuegeometric mean
Quartile/Quintile/Deciles/Percentile
Quantiles L=(n+1)y/100
The third quartile > mean 9-30
N n
X
N
MAD i 1
i X (X i ) 2
(X i X )2
For population: 2 i 1
For sample: s 2 i 1
n N n 1
C FA
Chebyshev
Inequality
Regardless of the shape of the distribution;
The proportion of the values that lie within k standard deviations of
F A
the mean is at least 1-1/k2
C
Standard deviation (risk) per unit of sample mean
s
CV= x
X
CV
Scale-free
Relative dispersion
R -R
Sharp ratio= P f
Sharpe excess return per unit of risk P
Ratio
10-30
Skewness ()
Fight fat tail
Positive
Mode<median<mean
skewed
Frequent small losses and a few extreme gains (mean=0)
prefer positive skewness
Left fat tail
Negative Mode>median> mean
C FA
Kurtosis ()
skewed Frequent small gains and a few extreme losses.(mean=0)
CF A
Sample kurtosis>3;Excess kurtosis>0
Leptokurtic
more frequent extremely large deviations from the
mean than a normal distribution
normal distribution
Investors dislike this distribution
11-30
R8 C FA
C F A
Probability Concepts
12-30
0 P(E) 1
Properties of
P(E1)+P(E2)++P(En)=1
Probability
E1En : mutually exclusive and Exhaustive
Empirical probability:
Priori probability:
C FA
Subjective probability:
F A
Mutually exclusive
C
P(AB)=P(A|B)=P(B|A)=0
P(AB)=P(A) P(B)
If exclusive , must not
Independence
independence
Independence =0,
13-30
Covariance & Correlation
How one random variable moves
with another random variable
Covariance COV E (X E(X))(Y E(Y)) The covariance of X with itself is
FA
equal to the variance of X
C
Covariance ranges from negative
infinity to positive infinity
(,)
relationship between tow random
variable
xy =
Correlation () Correlation has no units, ranges from
-1 to +1,standardization of covariance
If =0,this doesnt indicates
independence
14-30
Expected return and variance of a portfolio
n
1.
C FA
p 2 =
w1212+w2222+2w1w2121,2
2.
CF A
Variance = 1p= w11+w22p2
= -1p= [w11-w22]p2
p2
n
p2
ncovi,jp2
15-30
Bayes Formula ()
80%
30%
20%
C FA
10%
70%
A
90%
CF
P(A B)
P(B A)
P(A)
P(B)
17-30
Probability Distribution
The number of Measurable and
Discrete
outcomes is counted. positive probability
The number of P(x)=0 even though x
A
Continuous
outcomes is infinite.
C Fcan occur
Cumulative probability function F(x)=P(X<=x)
Discrete probability Distribution
CF A &
Discrete uniform
X 1,2,3,4,5,p(x) 0.2
1
Bernoulli random
P(Y=1)=p , P(Y=0)=1-p
variable
Expectation =p , Variance=p(1-p)
n
Binomial random p(x) P(X x) nCxp (1 p)
x n x
variable 18-30
Expectation =np , Variance = np(1-p)
Continuous Probability Distribution
&
Uniform (a , b)
Distribution P(x1 X x 2) (x 2 x1) / (b a)
X ~ N(, 2 )
Symmetrical distribution : skewness=0;kurtosis = 3
Properties A linear combination of normally distributed is also normally distributed
FA
The tails get thin and go to zero but extend infinitely
C
:(, +)
68% confidence interval ,
90% confidence interval 1.65, 1.65
A
Confidence
Normal
Distribution
intervals
C FA
Selected historical data to generate a distribution;
Historical simulation
F A
The past can not indicate the future;
C
Historical simulation cannot address the sort of what
if questions that Monte Carlo simulation can
20-30
R10 C FA
F A
Sampling and Estimation
C
21-30
Sampling
C FA
Data
Time-series data : data taken over a period of time
Cross-sectional data : data taken at a single point of time
CF A
Sampling error of the mean = sample mean-population mean
Sample statistic The sample statistic itself is a random variable
Central Limit Theory : n>=30sample mean ~ N(, 2 / n)
Standard error = / n or s / n
22-30
Estimation
Unbiasedness: expected value of the estimator is equal to the
Desirable parameter that are trying to estimate
properties Efficiency: dispersion is smaller
Consistency: the accuracy increases as n increases
Point Estimate:
Estimation Confidence
interval estimate
s
A
x z / 2 x t / 2
n or n
C F
zt
n>=30,z
CF A
Symmetrical: skewness = 0
Degrees of freedom (df): n-1
Less peaked than a normal distribution (fatter tails), kurtosis < 3
T
As the degrees of freedom gets larger, the shape of t-distribution approaches
standard normal distribution
significance leveltconfidence level
bias
Data-mining bias Look-ahead bias
Sample selection bias Time-period bias
23-30
Survivorship bias
R11 C FA
CF A
Hypothesis Testing
24-30
Two-tailed H0:=0 H0:0
1. One-tailed H0:0 H0 :0 or ,H0 :0 H0 :<0
H0 is what we want to reject
x 0
2.test statistic Test statistic=
s/ n
C FA
A
3.critical value
CF
Reject H0 if |test statistic|> critical value
*****is significantly different from *****
4.
Fail to reject H0 if |test statistic|< critical value
*****is not significant different from *****
25-30
Type 1 error and type 2 error
Decision H0 is true H0 is false
Incorrect Decision
Do not reject H0 Correct decision
Type 2 error
Incorrect decision Correct decision
Reject H0
Significance level =P(type 1 error)
C FA
Power of test =1-P(type 2 error)
Test type
Assumption H0 Test-statistic
Independent populations 1 2 0
Mean t
Paired comparisons test () d 0
Normally distributed population (1) 2 02 2
variance
Two independent population (2) 12 2 2 F text
26-30
Nonparametric tests ( )
C FA
When data are ranks (ordinal measurement scale)rather than
Nonparametric
tests
values
F A
The hypothesis does not involve the parameters of the
C
distribution , such as testing whether a variable is normally
distributed
27-30
R12 C FA
CF A
Technical Analysis
28-30
Technical analysis ()
Prices are determined by the interaction of supply and demand
Principles Only participants who actually trade affect prices
Price and volume reflect the collective behavior of buyers and sellers
Investor behavior is reflected in trends and patterns that trend to repeat
Assumptions
Efficient markets hypothesis does not hold
Fundamentalists believe that prices react quickly to changing stock values ,
A
while technicians believe that the reaction is slow
Actual price and volume data are observable
C F
Technical analysis itself is objective, while much of the data used in
A
fundamental analysis is subject to assumptions or restatements
Advantages
CF
It can be applied to the prices of assets that do not produce future cash flows ,
such as commodities
It can also be useful when financial statement fraud occurs
CFA CFA
C FA
CF A
CFA
30-30