Download as pdf or txt
Download as pdf or txt
You are on page 1of 42

Lecture 4

This lesson
• Feedback on classroom test 4
• Recap of CAPM
• Index models and factor models
• Exercise
Copyright © 2021 McGraw-Hill Education. All rights reserved. No reproduction or distribution wit
11-4

More on Correlation and the Risk-Return Trade-Off


(The Next Slide is an Excel Example)
From efficient frontier to CAPITAL MARKET LINE

E(rp)

M
CML Capital Market Line
Ü To price efficient portfolios
A
rf

sp Risk premium
E (rm ) - rf
E (rp ) = rf + ´s p
sm
6

Assumption of CAPM
• Individuals
– Mean-variance optimizers
– Homogeneous expectations
– All assets are publicly traded
• Markets
– All assets are publicly held
– All information is available
– No taxes
– No transaction costs
from CML to SML Video lezione 7

E (rm ) - rf s i,m s i,m


E (ri ) - rf = ´ = E (rm ) - rf ´ 2
sm sm sm
E (ri ) = rf + b i éë E (rm ) - rf ùû
SML Security Market
Line

E (rp ) = rf + b p éë E (rm ) - rf ùû
Security Market Line

SML
E[R(i)] = Rf +bi = [ R(m) – Rf] E (R)
Alto
ritorno
Pendenza SML
R(m) = =
m
Basso MRP
Market Risk
Premium (MRP) ritorno
Rf =

=
Basso
rischio
b(m) = 1 Alto
rischio
b

8
.

The Systematic Risk Principle

• What determines the size of the risk premium on a


risky asset/inefficient portfolios?

• The systematic risk principle states:

The expected return on an asset depends


only on its systematic risk.

• No matter how much total risk an asset has, only


the systematic portion is relevant in determining the
expected return (and risk premium) on that asset.
12-10

Measuring Systematic Risk

• To be compensated for risk, the risk has to be special.


– Unsystematic risk is not special.
– Systematic risk is special.

• The Beta coefficient (b) measures the relative systematic risk of an


asset.
– Assets with Betas larger than 1.0 have more systematic risk than average.
– Assets with Betas smaller than 1.0 have less systematic risk than average.

• Because assets with larger betas have greater systematic risks, they
will have greater expected returns.

Note that not all Betas are created equally.


12-11

Published Beta Coefficients


Beta by Bloomberg

• Enter this command to look up the beta of a stock:


– <ticker symbol> <EQUITY> BETA <GO>
• Example: The screenshot shows the result for the beta
of Goldman Sachs.
– GS <EQUITY> BETA <GO>
• Using the default settings, Bloomberg performs a
regression of the historical trading prices of the stock
against the S&P 500 (SPX) using weekly data over a two-
year period. Depending on the security, you can often
find data for the past 20-25 years!
Beta by Bloomberg

• Bloomberg reports both the Adjusted Beta and Raw


Beta. The adjusted beta is an estimate of a security's
future beta. It uses the historical data of the stock, but
assumes that a security’s beta moves toward the market
average over time. The formula is as follows:
• Adjusted beta = (.67) * Raw beta + (.33) * 1.0
12-14

Portfolio Betas

• The total risk of a portfolio has no simple relation to the total


risk of the assets in the portfolio.
– Recall the variance of a portfolio equation.
– For two assets, you need two variances and the covariance.
– For four assets, you need four variances and six covariances.

• In contrast, a portfolio Beta can be calculated just like the


expected return of a portfolio.

• That is, you can multiply each asset’s Beta by its portfolio
weight and then add the results to get the portfolio’s Beta.
Abnormal returns

E(Ri) C

A Aggressive securities
E(RA) D
M
E(RM)
F
E(RB B
)

E
Rf Defensive securities

1 (b)
12-25

Portfolio Expected Returns and Betas for two


Assets

Is this situation suitable in the CAPM world?


12-26

The Fundamental Result, I.

• The situation we have described for assets A and B


cannot persist in a well-organized, active market.
– Investors will be attracted to asset A (and buy A shares).
– Investors will shy away from asset B (and sell B shares).

• This buying and selling will make:


– The price of A shares increase.
– The price of B shares decrease.

• This price adjustment continues until the two assets plot


on exactly the same line.

E(R A ) - R f E(RB ) - R f
That is, until : =
βA βB
12-27

The Fundamental Result, II.

In general …

• The reward-to-risk ratio must be the same for all assets


in a competitive financial market.

• If one asset has twice as much systematic risk as another


asset, its risk premium will simply be twice as large.

• Because the reward-to-risk ratio must be the same, all


assets in the market must plot on the same line.
12-28

The Fundamental Result, III.


29

CAPM: take home messages

– CAPM is a model, built on hypotheses


– Know the definition of the market portfolio, beta, CML,
SML.
– Why all investors would hold the market portfolio
– Expected return-beta relationship
30

Exercise time
Suppose that you want to identify mispriced securities. You are considering to invest
in the following two stocks. Based on current dividend yields and expected capital
gains, your finance department gives you the following information:
• Stock ABC: E(R) = 12%; σ= 20.5%; β = 0.8;
• Stock XYZ: E(R) = 14.5%; σ = 38%; β = 1.35;

Moreover, from your finance department, you have obtained the following
information:
• S&P 500: E(R) = 13%; σ = 25%;
• The T-bill rate is currently 6%. (Risk free)

• Question 1. Suppose that you are currently holding the market-index portfolio.
According to the CAPM, would you choose to add either of these two stocks to
your holdings?
• Question 2. If you could invest only in T-bills and one of these portfolios, which
would you choose? Would your answer differ if the standard deviation of stock
ABC was 24%?
31

Exercise time
• Stock ABC: E(R) = 12%; σ= 20.5%; β = 0:8;
• Stock XYZ: E(R) = 14.5%; σ = 38%; β = 1.35;
• S&P 500: E(R) = 13%; σ = 25%;
• The T-bill rate is currently 6%. (Risk free)
• Question 1. Suppose that you are currently holding the market-index portfolio.
According to the CAPM, would you choose to add either of these two stocks to
your holdings?

Using the SML equation:


E(r )= rf+ β(MRP)
• Stock ABC: 6%+0.8*(13%-6%)=11.6%<12%
• Stock XYZ: 6%+1.35*(13%-6%)=15.45%>14.5%

Hence, Stock ABC is desirable


32

Exercise time
• Stock ABC: E(R) = 12%; σ= 20.5%; β = 0:8;
• Stock XYZ: E(R) = 14.5%; σ = 38%; β = 1.35;
• S&P 500: E(R) = 13%; σ = 25%;
• The T-bill rate is currently 6%. (Risk free)
• Question 2. If you could invest only in T-bills and one of these portfolios, which
would you choose? Would your answer differ if the standard deviation of stock
ABC was 24%?
The question asks which risky asset you would hold together with the risk-free
rate. This means that we need to compute the Sharpe ratio of our stocks (reward to
risk ratio).
Using the Sharpe ratio formula:
Sharpe ratio (A)=(12%- 6%)/20.5%= 0.2927
Sharpe ratio (B)=(14.5% -6%)/38%= 0.2237
Sharpe ratio (Market index portfolio)=(13% -6%)/25%= 0.28
So, Asset A would be a good substitute for the market-index portfolio.
Sharpe ratio (A’)=(12%- 6%)/24%= 0.25
33

CAPM in the real world:


the Index model
An Index Model is a Statistical model of security returns (as
opposed to an economic, equilibrium-based model).

A Single Index Model specifies two sources of uncertainty


for a security’s return:
Systematic (macroeconomic) uncertainty Unique (microeconomic) uncertainty
(which is assumed to be well represented (which is represented by a security-specific
by a single index of stock returns) random component)
Formalizing the Basic Idea:
The Return Generating Model
Index model as Market model

How to do estimate the equation?

Collect historical data on rj and rM


Run a simple linear regression of rj against rM

alpha” is the intercept


Beta is the slope
ej is regression error

The fitted regression line is called the Security Characteristic Line (SCL)
Two approaches to estimate SIM:

Subscript GM refers to stock GM (general motors, an example)


38

Security Characteristic Line

Excess Returns (i)


SCL

. . . .. .
. . . . ..
. .. . . . . . .
. . .
. . . . . .
. .. . . . . Excess returns

. . . ..
on market index
.
. . . .. . . .
R i = a i + ß i R m + ei
39

How to cast the index model?

Standard procedure: OLS regression

From theory:
CAPM: E(ri )= rf + b [E(Rm)]

Using historical data:


Regression eq.: Ri = α+ β Rm+ ei
40

The beta from the regression


Standard procedure

1. Compute rate of return from the data Ü


Lenght of the dataset?
It depends. Rule of thumb:
One-year of daily obs
Two years of weekly obs.
Five years of Monthly obs.

2. Compute the rate of return of a market index (using the same


frequency of observations!)
-> Select a market index!
41

Beta Estimation: The Index Effect


42
From the regression line

Regression equation: Ri = α+ β Rm+ ei

1. The α Ü When Rm =0, Ri = α (The stock’s expected return


beyond that induced by the market index)
2. The slope β Ü The sensitivity of a security’s returns to the market
factor
3. Epsilon (e) Ü Also called residuals in regression, and for securities,
residual returns.
- From a graphical perspective, they are the distance between the
point observation and the regression line
- They are a proxy for unexpected events that are relevant only to
this security (so…?)
43

The regression line

Ri = a + b Rm + e

b
44
From the regression line (Cont’d)

Jensen’s a = 𝛼! = 𝑟" − {𝑟# + 𝐸 𝑟$ − 𝑟# 𝜷𝒑 }


𝛽" can be estimated using the index model
𝑟" − 𝑟# can be replaced by the average risk premium of P (or the differences
between the 2 averages, same thing)
𝐸 𝑟$ − 𝑟# can be replaced by the average risk premium of the market (or the
differences between the 2 averages, same thing)

R2
- How much the model explains the data: in other words, how much the
portfolio’s movement can be explained by the market
- From 0 to 1
- What is (1-R2) ?

S.E.
- «error» on the beta estimation.
- Used to address the significance of the Beta, so the significance of our
estimation
45

Risk components from the regression


Ri = α+ β Rm+ ei

si2 = bi2 sm2 + s2(ei)


where;
si2 = total variance
bi2 sm2 = systematic variance
s2(ei) = unsystematic variance
46

Risk components (cont’d)

Total Risk = Systematic Risk + Unsystematic


Risk

Systematic Risk/Total Risk = R2

ßi2 s m2 / s2 = R2
47

From CAPM to
multifactor models
Multifactor models

• Use more than one factor in addition to market


return
– Examples include gross domestic product,
expected inflation, interest rates, etc.
– Estimate a beta or factor loading for each
factor using multiple regression
49

Fama and French three factor model


• Motivation:
• From late 80’s some Observations from the data:
Positive/negative alphas, under/overestimation of
parameters, instability over time

• Average returns on stocks of Small Firms and firms with high


ratios of book value of equity to market value of equity have
been higher than the ones predicted by the CAPM

• Multiple sources of Systematic risk?


50

Fama and French three factor model


• There are variables that have a strong role in explaining
average stock returns (size, book-to-market ratio, earning-
price ratio).
• If assets are priced rationally, this means that stock risks are
multidimensional.
• Previously: only the market index in the regression
– So we link the risk/return of a security only to the market
index movements
• Include in a multivariate analysis the size premium (SMB) and
the B/M premium (HML)
51

Fama and French three factor model


• It is possible to construct portfolios that track size and B/M
factors, and to compute the respective premiums.
• The size premium is computed as the difference in return
between small and large firms (SMB). The B/M premium is
computed as the difference in returns between firm with a
high versus low B/M ratio (HML).
• The new regression equation takes this form:
• 𝑟! − 𝑟" = 𝛼! + 𝛽 𝑟# − 𝑟" + 𝛽$%& 𝑟$%& + 𝛽'%( 𝑟'%(
Exercise time

• Exercise
– Use GameStop

You might also like