Lecture 2.

Quantitative Economics & Finance Analysis

Statistics Basics:
Normal
Distribution
Statistics Basics: • Definition of Arithmetic Mean
Mean • The arithmetic mean is the sum of the
observations divided by the number of
observations.

• Sample Mean Formula. The sample mean or

average, X (read ‘‘X-bar’’), is the arithmetic
mean value of a sample:
Statistics Basics: • Definition of Median
Median • The median is the value of the middle item of a
set of items that has been sorted into
ascending or descending order.

• In an odd-numbered sample of n items, the

median occupies the (n + 1)/2 position.

• In an even-numbered sample, we define the

median as the mean of the values of items
occupying the n/2 and (n + 2)/2 positions.
Statistics Basics: • Definition of Mode
Mode • The mode is the most frequently occurring
value in a distribution.
Statistics Basics:
Standard • Sample Standard Deviation
Deviation • We can compute a sample standard deviation
(Sigma) by taking the positive square root of the
sample variance.
Historical average
returns and
standard
deviation
Statistics Basics: The normal distribution
Normal
Distribution
Statistics Basics: • Sample Covariance
Covariance •The covariance between two random variables
is a statistical measure of the degree to which
the two variables move together.

•The covariance captures the linear relationship

between two variables. A positive covariance
indicates that the variables tend to move
together; a negative covariance indicates that
the variables tend to move in opposite
directions.
 • Sample Correlation Coefficient (Rho)
• The correlation coefficient, r, is a measure of the
Statistics Basics: strength of the linear relationship (correlation)
Correlation between two variables. The correlation coefficient
has no unit of measurement; it is a “pure” measure
of the tendency of two variables to move together.
• The sample correlation coefficient for two
variables, X and Y, is calculated as:

• The correlation coefficient is bounded by positive

and negative one (i.e., –1 ≤ r ≤ +1), where a
correlation coefficient of +1 indicates that changes in
the variables are perfectly positively correlated (i.e.,
they go up and down together, in lock-step). In
contrast, if the correlation coefficient is –1, the
changes in the variables are perfectly negatively
correlated.
Statistics Basics: • Portfolio Expected Return
Portfolio • Weighted average of the individual risky assets
Expected Return expected returns in the portfolio.
𝑁𝑁

𝐸𝐸 𝑅𝑅𝑝𝑝 = � 𝑤𝑤𝑗𝑗 𝐸𝐸 𝑅𝑅𝑗𝑗

𝑗𝑗=1

𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑗𝑗
• Where 𝑤𝑤𝑗𝑗 = ; ∑ 𝑤𝑤𝑗𝑗 = 1
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴
Example: A B C D E F G H
Portfolio 1 Stock V Stock W Stock X Stock Y Stock Z Portfolio
2 w(j) 30% 17% 22% 20% 12% 100%
Expected Return 3 State (i) p(i) Expected Return
4 Recession 0.25 -20.0% 18.0% 5.0% -8.0% 4.0% -3%
5 Neutral 0.50 17.5% 15.0% 10.0% 11.0% 9.0% 13%
6 Boom 0.25 35.0% -10.0% 15.0% 16.0% 12.0% 17%
7 E(R) 1.00 12.5% 9.5% 10.0% 7.5% 8.5% 10%

Steps:
5
1. Calculate expected portfolio return in E ( R P ,i ) = ∑ w j E ( R j )
each state: j =1

2. Apply the probabilities of each state to

3
the expected return of the portfolio in E ( R P ) = ∑ p i E ( R P ,i )
that state. i =1

3. Sum the result of step 2.

Statistics Basics: • Portfolio Total Risk
Portfolio Total • Portfolio with 2-stock, A & B. What is the portfolio
Risk Measure total risk?
• The total risk is measured by the standard deviation
𝜎𝜎.
• Thus, the portfolio standard deviation with 2-stock
is:
𝜎𝜎 𝑅𝑅𝑝𝑝 = 𝑤𝑤𝐴𝐴2 𝜎𝜎𝐴𝐴2 + 𝑤𝑤𝐵𝐵2 𝜎𝜎𝐵𝐵2 + 2𝑤𝑤𝐴𝐴 𝑤𝑤𝐵𝐵 𝐶𝐶𝐶𝐶𝐶𝐶 𝑅𝑅𝐴𝐴 , 𝑅𝑅𝐵𝐵

𝜎𝜎 𝑅𝑅𝑃𝑃 = 𝑤𝑤𝐴𝐴2 𝜎𝜎𝐴𝐴2 + 𝑤𝑤𝐵𝐵2 𝜎𝜎𝐵𝐵2 + 2𝑤𝑤𝐴𝐴 𝑤𝑤𝐵𝐵 𝜎𝜎𝐴𝐴 𝜎𝜎𝐵𝐵 𝜌𝜌𝐴𝐴𝐴𝐴

• Rho (r) or 𝜌𝜌, Correlation Coefficient, plays an

important to lower the portfolio risk.
Example: Portfolio

Portfolio Standard State (i) p(i) E( R )

Dev Dev^2 x p(i)

Deviation Recession 0.25 -3% -13% 0.01663 0.00416

Neutral 0.50 13% 3% 0.00101 0.00050
Boom 0.25 17% 7% 0.00428 0.00107
E(R) 1.00 10% VAR(Pf) 0.00573259
Std(Pf) 7.6%

1. Calculate expected portfolio return in each state of the economy and

overall.
2. Compute deviation (DEV) of expected portfolio return in each state from
total expected portfolio return.
3. Square deviations (DEV^2) found in step 2.
4. Multiply squared deviations from step 3 times the probability of each state
occurring (x p(i)).
5. The sum of the results from step 4 = portfolio variance.
 𝜎𝜎
Regression: Unsystematic Risk or
Diversifiable Risk
CAPITAL ASSET
PRICING (CAPM)

Capital Asset Pricing Model (CAPM)

• The relationship between risk and return. Systematic Risk
• Only systematic or market risk is priced in the model. Or Non-diversifiable Risk

Investors are compensated for risk which is non-

diversifiable.
𝐸𝐸 𝑅𝑅𝑖𝑖𝑖𝑖 = 𝑅𝑅𝑓𝑓 + 𝛽𝛽 𝑅𝑅𝑚𝑚𝑚𝑚 − 𝑅𝑅𝑓𝑓
 If CAPM is true to price the risky assets, the expected
E(Rit)
return of the risky assets should fall on SML.
Regression: SML
Security Market Fair price or #
Line (SML) Intrinsic value Undervalue
𝐸𝐸 𝑅𝑅𝑖𝑖𝑖𝑖 .
Mispricing in market

Expected return of risky asset *

Overvalue
changes if:
𝑅𝑅𝑚𝑚 − 𝑅𝑅𝑓𝑓
𝑅𝑅𝑓𝑓
1. 𝑅𝑅𝑓𝑓 : Risk-free rate
2. 𝑅𝑅𝑚𝑚 − 𝑅𝑅𝑓𝑓 : Market risk premium
3. 𝛽𝛽: beta the sensitivity to the
market 𝛽𝛽
𝛽𝛽
 Apply CAPM for Investment
Regression: 𝐸𝐸 𝑅𝑅𝑖𝑖𝑖𝑖 = 𝑅𝑅𝑓𝑓 + 𝛽𝛽 𝑅𝑅𝑚𝑚𝑚𝑚 − 𝑅𝑅𝑓𝑓 ------------------------- (A)
CAPM in Practice Steps:
1. Estimate risk-free rate using treasury yield, i.e. 3-month, 6-month,
12-month, 2-year, or 5-year. Other risk-free rates are possible if it
explanatory power is high in the model.
2. Estimate market return using market index, i.e. STI, ASX50, S&P500,
etc. to proxy for the market return. Specifically,
𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑡𝑡 − 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑡𝑡−1
6. Review and rebalance your portfolio. 𝑅𝑅𝑚𝑚 =
𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝑡𝑡−1
3. Estimate beta 𝛽𝛽 using (1) the characteristic line for regression over
a period of time t; (2) Statistics measure for beta. Specifically,
(1) 𝑅𝑅𝑖𝑖𝑖𝑖 = 𝛼𝛼 + 𝛽𝛽𝑅𝑅𝑚𝑚𝑚𝑚
𝐶𝐶𝐶𝐶𝐶𝐶𝑅𝑅𝑖𝑖 ,𝑅𝑅𝑀𝑀
(2) 𝛽𝛽 =
𝑉𝑉𝑉𝑉𝑉𝑉(𝑅𝑅𝑀𝑀 )
4. Estimate CAPM required rate of return in equation (A).
5. Compare Step 4 with the security’s expected return in the stock
exchange to identify under- and over-valued stocks to form a long-
short portfolio.
 Regression:
CAPM in
Research- Proxy
for 𝑅𝑅𝑀𝑀

Exhibit 8.18 illustrates the effects of two market proxies on

betas of the 30 DJIA stocks over three different periods
Exhibit 8.19 contains the risk-free rates, market returns, the
slopes of the SMLs

The beta intercept of the SML will differ if:

 There is an error in selecting the risk-
free asset.
 There is an error in selecting the market
portfolio.
 Regression: Fama and French 3-Factor Model
Fama & French
• Developed by Fama and French (1992).
3-Factor Model
• expands on the capital asset pricing model (CAPM) by
adding size risk and value risk factors to the market
risk factor in CAPM.
Small-R H-L
𝐸𝐸 𝑅𝑅𝑖𝑖𝑖𝑖 = 𝑅𝑅𝑓𝑓 + 𝛽𝛽1 𝑅𝑅𝑚𝑚𝑚𝑚 − 𝑅𝑅𝑓𝑓 + 𝛽𝛽2 𝑆𝑆𝑆𝑆𝑆𝑆𝑡𝑡 + 𝛽𝛽3 𝐻𝐻𝐻𝐻𝐻𝐻𝑡𝑡
 Please visit Prof. Kenneth French’s Faktor 1 Faktor 2 Faktor 3
website for the data for FF-3 Factor
Where,
model.
https://mba.tuck.dartmouth.edu/page 𝑆𝑆𝑆𝑆𝑆𝑆𝑡𝑡 is size premium (small minus big)
s/faculty/ken.french/data_library.html 𝐻𝐻𝐻𝐻𝐻𝐻𝑡𝑡 is value premium (high minus low)
 After knowing how to apply CAPM,
are you able to apply FF-3 Factor
model for your investment
strategy?
 • The APT Model
Regression:
Arbitrage Pricing E(Ri)=λ0+ λ1bi1+ λ2bi2+…+ λkbik
Theory (APT) where:

Model λ0=the expected return on an asset with zero

systematic risk
λj=the risk premium related to the j th common risk factor
bij=the pricing relationship between the risk premium and the asset;
that is, how responsive asset i is to the j th common factor

• Security return are governed by a set of

broad economic influences in the following
fashion by Chen, Roll, and Ross in 1986.

Rit = ai + [bi1Rmt + bi 2 MPt + bi 3 DEIt + bi 4UIt + bi 5UPRt + bi 6UTSt ] + eit

where:
Rm= the return on a value weighted index of NYSE-listed stocks UI=the difference between actual and expected levels of inflation
MP=the monthly growth rate in US industrial production UPR=the unanticipated change in the bond credit spread
DEI=the change in inflation, measured by the US consumer price index UTS= the unanticipated term structure shift (long term less short term RFR)
Portfolio • Treynor ‘s Composite Performance Measure
Performance:
Treynor Ratio

T=
(R i − RFR )
βi
• The numerator is the risk premium
• The denominator is a measure of risk
• The expression is the risk premium return per unit of
risk
• Risk averse investors prefer to maximize this value
• This assumes a completely diversified portfolio
leaving systematic risk as the relevant risk
Portfolio • Risk premium earned per unit of risk
Performance:
Sharpe Ratio

R i − RFR
Si =
σi
 Portfolio • Demonstration of Comparative Sharpe Measure
Performance:
Sharpe Ratio Portfolio
Average Annual Rate of Standard Deviation of
Return Return
D 0.13 0.18
E 0.17 0.22
F 0.16 0.23

0.14 − 0.08 0.13 − 0.08

=SM = 0.300 =SD = 0.278
0.20 0.18

0.17 − 0.08 0.16 − 0.08

=SE = 0.409 =SF = 0.348
0.22 0.23

•The D portfolio had the lowest risk premium return per unit of total risk, failing even to
perform as well as the aggregate market portfolio. In contrast, Portfolio E and F performed
better than the aggregate market: Portfolio E did better than Portfolio F.
Treynor v.s.
Sharpe Treynor versus Sharpe Measure

• Sharpe uses standard deviation of returns as the measure of

risk
• Treynor measure uses beta (systematic risk)
• Sharpe therefore evaluates the portfolio manager on the
basis of both rate of return performance and diversification
• The methods agree on rankings of completely diversified
portfolios
• Produce relative not absolute rankings of performance
 Jensen Portfolio
Jensen’s Alpha Performance Measure
Also based on CAPM
Expected return on any security or portfolio is

E (R j ) = RFR + β j [E(R m ) − RFR ]

 Jensen Measure and Multifactor Models
Jensen’s Alpha
• Advantages:
• It is easier to interpret
• Because it is estimated from a regression equation, it
is possible to make statements about the statistical
significance of the manger’s skill level
• It is flexible enough to allow for alternative models of
risk and expected return than the CAPAM. Risk-
adjusted performance can be computed relative to
any of the multifactor models:

α j + [b j1 F1t + b j 2 F2t + b jk Fkt ] + e jt

R jt − RFRt =