You are on page 1of 52

COMM371/COEC371 – Investment Theory

Lecture 5
Capital Asset Pricing Model (CAPM)

Instructors

Prof.LOGO
Alberto Mokak Teguia
VERSIONS
The UBC Sauder logo should always appear in its full colour version. The door frame is in green and the door
wordmark are in blue. However, there may be special circumstances where the logo has to appear on a colou
Jose Pizarro
background, in which case:

When the background is light, the logo should appear in the preferred full colour; however, there must be sig
contrast between the background and both the blue and green of the logo. Using grey as an example, the bac
Dongliang Lu
must be no darker than 15% black to use the full colour logo.

FULL COLOUR ON WHITE FULL COLOUR ON LIGHT


BACKGROUND (15% BLACK)

When the background is black or dark, the logo reverses to white.


1/52
COMM371/COEC371 – Investment Theory

Question: What Determines Asset Returns?

Historical data suggest that different assets offer different average


returns
– Stock portfolios perform better than bond portfolios
– Small cap stocks perform better than large cap stocks

Risk-return trade-off: return is seen as compensation for risk taken

What risk is compensated?

How is the trade-off quantified?

2/52
COMM371/COEC371 – Investment Theory

Answer: CAPM

The Capital Asset Pricing Model (CAPM) provides answers to


the above questions, assuming the market is in equilibrium. In
particular,
– How are expected returns determined?

– How are they related to risk?

It is theoretical model which is not perfect but it serves well as a


starting point.

3/52
COMM371/COEC371 – Investment Theory

Overview of the Lecture

1. From Portfolio Theory to the CAPM

2. CAPM: Assumptions and Implications

3. The CAPM Equation

4. SML and CML

5. Beta and Alpha

6. Applications of the CAPM

7. Statistical Tests of the CAPM

8. Alternative Models to the CAPM

4/52
COMM371/COEC371 – Investment Theory

From Portfolio Theory to an Equilibrium Model


Portfolio Theory
– Optimal individual behavior (portfolio optimization) given returns
– No market-wide considerations

Capital Asset Pricing Model


– Optimal individual behavior (portfolio optimization)
– Market-wide equilibrium determines expected returns

Why do we need an equilibrium model?


– Suppose an asset offers a stream of risky cash flows in the future.
Today’s price and expected future cash flows will determine the as-
set’s expected return.
– How much are investors willing to pay for this asset today?
– This important question cannot be answered without a wide-market
equilibrium model.
5/52
COMM371/COEC371 – Investment Theory

From Portfolio Theory to the CAPM

Expected returns of assets are not exogenously given - they are


determined in equilibrium

Equilibrium is the outcome of actions of rational agents in com-


petitive markets

Demand and supply in the market determine the price of risk, i.e.,
expected returns of assets

CAPM (Capital Asset Pricing Model) is a theoretical model which


gives a set of predictions concerning equilibrium expected returns
on risky assets, derived using principles of diversification and sim-
plifying assumptions.

6/52
COMM371/COEC371 – Investment Theory

The CAPM: Assumptions

Investors cannot affect prices by their individual trades (price


takers)

Investors have the same holding period (investment horizon)

There are multiple publicly traded risky assets and one risk-free
asset

Investors have the same information (beliefs)

Investors are rational mean-variance optimizers

No short sales and no frictions such as transaction costs and


taxes

7/52
COMM371/COEC371 – Investment Theory

The CAPM: Main Implications

1. All investors hold the same portfolio of risky assets - tangent port-
folio.

2. Tangent portfolio contains all risky assets, i.e., it coincides with


the market portfolio.

3. Market risk premium, E[RM ] − Rf , is determined by the average


risk aversion of all market participants.

4. Risk premium on an individual asset is proportionate to its covari-


ance with the market portfolio: E[Ri ] − Rf = λCov [Ri , RM ].

8/52
COMM371/COEC371 – Investment Theory

Tangent Portfolio

1. All investors hold the same portfolio of risky assets.

Since all investors have the same holding period, care only about
mean and variance, and use the same information, then they
should arrive to the same optimal portfolio of risky assets.

This is the Tangent Portfolio.

Investors differ only in the fraction of their complete portfolio in-


vested in the Tangent Portfolio.

9/52
COMM371/COEC371 – Investment Theory

Tangent Portfolio, cont’d


2. Tangent Portfolio contains all risky assets and the weights of as-
sets are equal to the ratios of the market value of each asset to
the total market value of all assets.

In equilibrium, the aggregate asset demand has to be equal to the


aggregate asset supply.
– Asset supply is all risky assets traded in the market, i.e., the market
portfolio.
– Asset demand is a sum of all investors’ risky portfolios, i.e., the tan-
gent portfolio.

So the Tangent Portfolio is also called the Market Portfolio.

Mutual Fund Theorem: Only one risky portfolio (market portfolio)


is needed by investors.

All investors invest in combinations of the Market Portfolio and the


risk-free asset.
10/52
COMM371/COEC371 – Investment Theory

The Market Portfolio (Value-Weighted)

The optimal tangent portfolio consists of all risky assets:

– N = number of risky assets in the economy


– Vi = market value of asset i (number of shares × price)
– VM = value of the Market Portfolio
– VM = V1 + V2 + · · · + VN
– wiM = weight of asset i in Market Portfolio = Vi /VM
PN M PN
– i=1 wi = i=1 Vi /VM = 1

Since the Market Portfolio includes all assets, each investor


holds all risky assets

11/52
COMM371/COEC371 – Investment Theory

Market Price of Risk in Equilibrium

3. Market risk premium, E[RM ] − Rf , is determined by the average


risk aversion of all market participants.

If investors become more risk averse, then


(1) They would hold more of the risk-free asset
(2) They would hold more of the market portfolio they should be
compensated with higher expected market portfolio return.

If investors are almost risk neutral, the market risk premium E[RM ]−
Rf would be close to zero.

12/52
COMM371/COEC371 – Investment Theory

Expected Return on Individual Assets


4. Risk premium on an individual asset is proportionate to its
covariance with the market portfolio: E[Ri ] − Rf = λCov [Ri , RM ].
Since the tangent portfolio is the Market Portfolio, the risk that
matters is Market Risk.

In equilibrium, the expected return on an individual asset should


depend on its contribution to overall risk, that is Market Risk.

Market Risk, captured by Var [RM ], is a weighted average of co-


variances of all asset returns with the Market Portfolio return:

Var [RM ] = Cov [RM , RM ]


X 
N
M
= Cov wi Ri , RM
i=1
XN
= wiM Cov [Ri , RM ]
i=1

13/52
COMM371/COEC371 – Investment Theory

Expected Return on Individual Assets, cont’d


It is reasonable to expect that, for an asset, the higher the covari-
ance with the market, the higher the risk premium.
Let us guess the following relation E[Ri ] − Rf = λCov [Ri , RM ] and
try to determine the constant λ.
Market Return is a weighted average of returns of all the assets:
XN
RM = wiM Ri
i=1
XN
RM − Rf = wiM (Ri − Rf )
i=1
XN
E[RM ] − Rf = wiM (E[Ri ] − Rf )
i=1
XN
E[RM ] − Rf = λ wiM Cov [Ri , RM ]
i=1
E[RM ] − Rf = λVar [RM ]

E[RM ] − Rf
⇒ λ=
Var [RM ]
14/52
COMM371/COEC371 – Investment Theory

The CAPM Equation


E[RM ]−Rf
Using λ = Var [RM ] we can write

E[RM ] − Rf Cov [Ri , RM ]


E[Ri ] − Rf = Cov [Ri , RM ] = (E[RM ] − Rf )
Var [RM ] Var [RM ]
Or

E[Ri ] − Rf = βi (E[RM ] − Rf )

where

Cov [Ri , RM ]
βi =
Var [RM ]

βi measures the extent to which return on asset i responds to


changes in the market portfolio returns.
βi is the regression slope coefficient of the asset i excess return
on the market portfolio excess return.
15/52
COMM371/COEC371 – Investment Theory

William Sharpe on the CAPM1

Portfolio theory focused on the actions of a single investor. I


said, What if everyone was optimizing? Then some people
decide they want to hold more IBM, but there aren’t enough
shares to satisfy demand. So they put price pressure on IBM
and up it goes, at which point they have to change their es-
timate of risk and return, because now they’re paying more
for the stock. That process of upward and downward pres-
sure on prices continues until prices reach an equilibrium and
everyone collectively wants to hold what’s available. At that
point, what can you say about the relationship between risk
and return? The answer is that expected return is proportion-
ate to beta relative to the market portfolio.

1 Willam Sharpe received the Nobel Prize in 1990 for his development of the CAPM.

Here are his comments from a 1998 interview with Jonathan Burton: “Revisiting the
Capital Asset Pricing Model,” Dow Jones Asset Manager (May/June 1998): 20–28.
16/52
COMM371/COEC371 – Investment Theory

Security Market Line (SML)


Graphical representation of the expected return and beta relation-
ship
E[Ri ] − Rf = βi (E[RM ] − Rf )
where
Cov [Ri , RM ]
βi =
Var [RM ]

Note: SML plots individual asset expected return as a function of


asset risk represented by beta.
17/52
COMM371/COEC371 – Investment Theory

CML and the Market Price of Risk


Capital Market Line (CML)

Slope of the CML is the market’s risk premium per unit of standard
deviation (market price of risk)
Slope on the CML is proportional to the average risk aversion of
investors in the economy
18/52
COMM371/COEC371 – Investment Theory

SML vs. CML

The SML

– is in the beta/expected return space


– contains all portfolios (according to the CAPM)

The CML

– is in the standard deviation/expected return space


– contains only the frontier portfolios

19/52
COMM371/COEC371 – Investment Theory

Beta
The beta on the market portfolio is equal to one, i.e., βM = 1.

Recall that β is a measure of sensitivity to changes in market


portfolio return.

If the covariance of an asset return with the market portfolio return


is positive then β > 0 and the expected return on the asset is
higher than Rf .

If the covariance of an asset return with the market portfolio return


is negative then β < 0 and the expected return on the asset is
lower than Rf .

β > 1 means that the asset is more risky than the market portfolio
and has a higher expected return.

β < 1 means that the asset is less risky than the market portfolio
and has a lower expected return.

20/52
COMM371/COEC371 – Investment Theory

β for Portfolios
The CAPM equation applies not just to individual assets but also
to portfolios:

E[Ri ] − Rf = βi (E[RM ] − Rf ) ⇒
wi (E[Ri ] − Rf ) = wi βi (E[RM ] − Rf ) ⇒
XN
E[Rp ] − Rf = wi (E[Ri ] − Rf )
i=1
X 
N
= wi βi (E[RM ] − Rf )
i=1

So for a portfolio of risky assets:

E[Rp ] − Rf = βp (E[RM ] − Rf )

where
XN
βp = wi β i
i=1

21/52
COMM371/COEC371 – Investment Theory

Example: Portfolio Beta

Rf = 1%, E[RM ] = 20%


Three assets A, B, and C: βA = 1.2, βB = 0.5, βC = −0.2.
Portfolio with wA = 0.5, wB = 0.3, wC = 0.2
What is the expected return on the portfolio, assuming that the
CAPM holds?
Answer: RP = 14.49%

Method 1: Use CAPM equation to find expected returns on A, B,


and C and then use the given weights to find the expected return
on the portfolio.

Method 2: Find portfolio beta and then use the CAPM equation
for the portfolio.

22/52
COMM371/COEC371 – Investment Theory

Alpha

Alpha (α) is the difference between an asset’s expected return


and the expected return predicted by the CAPM equation.

αi = (E[Ri ] − Rf ) − βi (E[RM ] − Rf )

If the CAPM holds then


– α > 0 implies that the asset is underpriced
– α < 0 implies that the asset is overpriced

23/52
COMM371/COEC371 – Investment Theory

Alpha: Example
Suppose that E[RM ] = 20%, Rf = 1%, βi = 1.3, E[Ri ] = 35%.
According to the CAPM, the asset’s expected excess return is:
1.3 × (20% − 1%) = 24.7%
Therefore, αi = 9.3%

24/52
COMM371/COEC371 – Investment Theory

Problem #1:

Assume the risk-free rate is 5% and the expected rate of return


on the market is 13%.

A share of stock A sells now for $60 and it will pay a dividend of
$4 at the end of year. The beta of the stock is 1.2.

Assuming that the CAPM holds, what do investors expect the ex-
dividend price of the stock to be at the end of year?

If, according to your research, the expected ex-dividend price at


the end of the year is $68, what is the alpha of stock A in your
view?

25/52
COMM371/COEC371 – Investment Theory

Solution:

According to CAPM, the expected return on stock A should be

E[RA ] = Rf + βA (E[RM ] − Rf )
= 0.05 + 1.2 × (0.13 − 0.05) = 0.146

Denoting by E[P1 ] the expected price next year, we have

E[RA ] = (E[P1 ] + D1 − P0 )/P0 = (E[P1 ] + 4 − 60)/60.

But according to the CAPM, E[RA ] should be equal 0.146. This


means E[P1 ] can be determined by solving

(E[P1 ] + 4 − 60)/60 = 0.146 ⇒ E[P1 ] = 56 + 0.146 × 60 = 64.76.

So, if the CAPM holds, investors expect stock A to sell at $64.76.

26/52
COMM371/COEC371 – Investment Theory

Solution continued:

If the expected price at the end of the year is $68, then the expected
return is
68 + 4 − 60 12
E[RA ] = = = 0.20.
60 60
Therefore, the alpha of stock A is

αA = (E[RA ] − Rf ) − βA (E[RM ] − Rf )
= (0.20 − 0.05) − 1.2 × (0.13 − 0.05) = 5.4%

Alternatively, you can also get αA by:

αA = (68 − 64.76)/60 = 5.4%

27/52
COMM371/COEC371 – Investment Theory

Problem #2:

Consider two stocks: one aggressive (A) and one defensive (D).

In bad times, the expected market return is 3%, the expected re-
turn on A is 1%, and the expected return on D is 3.8%.

In good times, the expected market return is 18%, the expected


return on A is 31% and the expected return on D is 12.8%.

If the CAPM holds, what are the betas of stocks A and D (assum-
ing they are constant across cycles)?

28/52
COMM371/COEC371 – Investment Theory

Solution:

Stock A:
E[RA ] − Rf = βA (E[RM ] − Rf )

0.01 − Rf = βA (0.03 − Rf ) (Bad times)


0.31 − Rf = βA (0.18 − Rf ) (Good times)

Subtracting the first equation from the second, we get

0.30 = βA × 0.15 ⇒ βA = 2

29/52
COMM371/COEC371 – Investment Theory

Solution continued:

Stock D:
E[RD ] − Rf = βD (E[RM ] − Rf )

0.038 − Rf = βD (0.03 − Rf ) (Bad times)


0.128 − Rf = βD (0.18 − Rf ) (Good times)

Subtracting the first equation from the second, we get

0.09 = βD × 0.15 ⇒ βD = 0.6

30/52
COMM371/COEC371 – Investment Theory

Applications of the CAPM

Security Analysis
– Testing the efficient market hypothesis.
– Identify mispriced securities
– Positive alpha suggests the asset is underpriced
– Negative alpha suggests the asset is overpriced

Capital Budget Decisions


– Valuation of firm’s investments (Corporate Finance)
– CAPM provides a risk-adjusted discount rate for the Present Value
calculation
– Example: New project has estimated β = 1.3. If Rf = 4% and the
market risk premium is 8%, the expected rate of return (according
to the CAPM) should be 4% + 1.3 × 8% = 14.4%. We can use this
rate to calculate PV of cash flows for the new project.

31/52
COMM371/COEC371 – Investment Theory

Statistical Tests of the CAPM


CAPM is a theoretical model

How do we assess its validity?


We can test some of its main predictions:
– Expected excess returns are linear in beta
– Slope of the SML is market risk premium
– Expected returns depend only on beta

Two issues about testing the CAPM:


(1) It is based on the "market portfolio" which includes all assets

– Empirical approach: Use a comprehensive index portfolio (e.g.,


S&P 500) as a proxy to the true market portfolio.
(2) It prescribes a relationship of expected–not realized–returns.

– Approach: Employ a regression model that allows using realized


returns.
32/52
COMM371/COEC371 – Investment Theory

Linear Regression
Consider two random variables X and Y .
Fact: To capture the relationship between X and Y , we can write
Y = α + βX + ϵ

– X : regressor, explanatory variable, or independent variable


– Y : Outcome variable, or dependent variable
– α, β: constant parameters (called the intercept and slope, respec-
tively)
– ϵ: random variable, such that
Cov [X , ϵ] = 0 and E[ϵ] = 0.
Variation in Y is decomposed into
– βX : variation in Y “explained" by X
– ϵ: “unexplained" variation
Fact: β is given by
Cov [X , Y ]
β= .
Var [X ]
33/52
COMM371/COEC371 – Investment Theory

Linear Regression, cont’d

Equation
Y = α + βX + ϵ
is the (linear) regression equation.

It is a useful way of thinking about how Y is generated.

Mechanism:
– α and β are given.

– For each “observation", an X and an ϵ are randomly drawn.

– Y is then determined according to the regression equation.

34/52
COMM371/COEC371 – Investment Theory

Regression Estimation

In general, we do not know the parameters α (intercept) and β


(slope)

However, we can estimate them using regression

Regression:
– We consider the scatterplot of X vs. Y

– We fit the “best" line to the scatterplot

– Suppose that line is


Y = α̂ + β̂X .

– α̂ is the estimate of α, and β̂ is the estimate of β.

35/52
COMM371/COEC371 – Investment Theory

A Scatterplot
Scatterplot and Best Line Fitting the Data
0.6

0.5

0.4

0.3

0.2
Y

0.1

-0.1

-0.2

-0.3

-0.4
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6
X

36/52
COMM371/COEC371 – Investment Theory

CAPM Regression
When using regression analysis to test the CAPM, we assume
that
– X is excess return of market portfolio, RM − Rf ,

– Y is excess return of the i-th asset, Ri − Rf .

Regression equation is

Ri − Rf = αi + βi (RM − Rf ) + ϵi .

Variation in returns of the i-th asset is decomposed into


– βi (RM − Rf ): this term captures systematic risk, i.e., risk that is
perfectly correlated with the market portfolio return

– ϵi : this term captures idiosyncratic risk, i.e., risk that is specific to


asset i and uncorrelated with the market portfolio return
37/52
COMM371/COEC371 – Investment Theory

Asset Characteristics

Three characteristics of asset i captured in the CAPM regression:

Ri − Rf = αi + βi (RM − Rf ) + ϵi

αi (alpha): measure of mispricing with respect to the CAPM

βi (beta): measure of sensitivity to market-wide movements

σ[ϵi ]: measure of idiosyncratic risk

38/52
COMM371/COEC371 – Investment Theory

Regression Output
Coefficient Interpretation
Estimate of α, denoted by α̂, also called Ab- Measure attractiveness of a firm,
normal return which should be 0 if CAPM holds.
Standard error of α̂ (measure of precision of How confident you are in the esti-
α̂), denoted by sα̂ mate of α.
Estimate of β, denoted by β̂ The usual interpretation of β:
Cov [Ri , RM ]
β̂i =
Var [RM ]

Standard error of β̂ (measure of precision of How confident you are in the esti-
β̂), denoted by sβ̂ mate of β
How well future asset returns are
Explained Variation likely to be predicted by the market
R-Square:R 2 = returns.
Total Variation
Var [βX ]
=
Var [βX ] + Var [ϵ]
β 2 Var [X ]
= .
β 2 Var [X ]+ σ[ϵ]2

39/52
COMM371/COEC371 – Investment Theory

Regression Example

Apple Inc. vs. S&P500. Monthly returns 01/1990 - 03/2022.

40/52
COMM371/COEC371 – Investment Theory

Regression Example: Output


Coefficient Estimate
Estimate of α, denoted by α̂, also called Ab- 0.013 or 1.3%
normal return
Standard error of α̂ (measure of precision of 0.006 or 0.6%
α̂), denoted by sα̂
For H0: α̂ = 0, we can construct a t-statistic
1.3%
α̂ − 0 t= = 2.37
t= 0.6%
sα̂ with p-value = 0.018

Estimate of β, denoted by β̂ 1.288


Standard error of β̂ (measure of precision of 0.128
β̂), denoted by sβ̂
For H0: β̂ = 0, we can construct a t-statistic
1.288
β̂ − 0 t= = 10.090
t= 0.128
sβ̂ with p-value = 0.000

R-Square 0.209

41/52
COMM371/COEC371 – Investment Theory

Regression Example: Scatterplot

42/52
COMM371/COEC371 – Investment Theory

CAPM Implications and Summary of the Empirical


Evidence

The CAPM equation is

E(Ri ) − Rf = βi (E(RM ) − Rf )

Testable Implications:

– Expected excess returns are linear in beta

– Slope of the SML is market risk premium

– Expected returns depend only on beta

43/52
COMM371/COEC371 – Investment Theory

CAPM Implications and Summary of the Empirical


Evidence

Empirical evidence for portfolios formed on the basis of beta:


– Expected excess portfolio returns are approximately linear in beta

– The slope of the SML line is smaller than market risk premium.

Do expected returns depend on other factors, in addition to beta?


– They do not seem to depend on idiosyncratic risk
– They seem to depend on firm size
Expected returns of small stocks exceed those of large stocks
(holding beta equal)
– They seem to depend on “book-to-market", the ratio of firm book
value to market value
Expected returns of high book-to-market stocks exceed those of low
book-to market stocks (holding beta equal)

44/52
COMM371/COEC371 – Investment Theory

Alternative Models to the CAPM

Multifactor models

– Two-factor model

– Fama and French model

Arbitrage-based models

– Arbitrage Pricing Theory (APT)

45/52
COMM371/COEC371 – Investment Theory

Multifactor Models

CAPM provides a basic insight that risk is decomposed into sys-


tematic and idiosyncratic parts.

Market risk does not fully account for systematic risk.

It appears that other economy-wide risk factors contribute to sys-


tematic risk as well.

Expected returns should still be compensation to exposure to these


risk factors.

46/52
COMM371/COEC371 – Investment Theory

Two-Factor Model Example

The CAPM is a single-factor model.

Multifactor models are natural extensions of the CAPM because


they incorporate additional risk factors, i.e., business-cycle vari-
ables.

Example: Consider a pension fund managing a bond and stock


portfolio.

The pension fund is exposed to


– stock market risk

– interest rate risk

The pension fund may prefer stocks that have small beta with re-
spect to interest rates (holding all else equal).

47/52
COMM371/COEC371 – Investment Theory

Two-Factor Model Example, cont’d

Pension fund manager would prefer to select stocks based on a


two-factor model:

E[Ri ] − Rf = αi + βiM (E[RM ] − Rf ) + βiTB (E[RTB ] − Rf )

– It accounts for market risk and interest rate risk

– Risk factors: market portfolio return RM and T-bond return RTB

– βiM is sensitivity to market risk of asset i

– βiTB is sensitivity to interest rate risk of asset i

48/52
COMM371/COEC371 – Investment Theory

How to Select Factors?

Open issue: identification of the important (“priced") factors.

Intense ongoing academic research and debate.

Two main approaches:

(1) Identify major systematic risks on theoretical grounds. For instance,


shocks to macroeconomic variables such as industrial production,
inflation, interest rates

(2) Search for factors on grounds of empirical success. For instance,


Fama-French factors based on size and book-to-market character-
istics.

49/52
COMM371/COEC371 – Investment Theory

Commonly-used Factors

50/52
COMM371/COEC371 – Investment Theory

Smart-Beta Investment Products

Investors get return compensation for being exposed to different


risk factors.

However, it’s hard for individual investors to replicate these factors


on their own.

How can investors get exposure to these factors?

Smart-beta products - those provide investors exposure to risk


factors (other than the broad market risk) across asset classes
that have also been found to explain additional sources of returns.

51/52
COMM371/COEC371 – Investment Theory

Smart-Beta ETFs

Source: etf.com
52/52

You might also like