Capital Asset Pricing Model

(CAPM) for Dummies !!

- Hrishikesh Herekar

Abstract

A lot of material exists on the net related to the Capital Asset Pricing Model (CAPM)
which gives tons of information regarding the history, assumptions, complex
derivations, description and shortcomings of the model. This paper does not intend
to do any of that. The paper intends to explain in layman’s terms what CAPM is all
about. The paper is meant for a person generally interested in finance and who is
reading about the CAPM for the first time and has at least a preliminary knowledge of
statistics and finance. No specialized knowledge is required to read this paper. Once
one gets an idea of the model one can easily delve deeper into the details of the
model through the information available on the net.

Table of Contents

1.0 Introduction ..........................................................................................2

2.0 Logic behind the CAPM............................................................................2
3.0 The CAPM formula ..................................................................................2
4.0 The Beta of an Asset...............................................................................4
5.0 Conclusion.............................................................................................4
6.0 Material for Further Reading ....................................................................4
7.0 Appendix I – The Portfolio Theory Derivation .............................................5

Dec 2006 Page 1 of 6 Hrishikesh Herekar

1.0 Introduction

Those of you, who trade in securities typically equities, must have heard a number of
times that a share is overpriced or under priced. Obviously this overpricing or under
pricing has to be with reference to “something”. This “something” is the “fair price”
of the security at which it gives an expected return which is commensurate with the
risk that the security inherently has. For example it is common sense that a risk free
security (US government bonds are assumed to risk free) is expected to give lower
returns as compared to a risky asset (Say an equity share of a company).
Unfortunately there is no one in this world who knows what the “fair price” of a
security really is. There are different valuation methods which try to find this, but
there is a lot of ambiguity involved.

CAPM is one of the models which try to price this risk. CAPM states that if certain
assumptions are accepted as given, then the expected return from an asset can be
calculated from the model. The following material tries to explain the CAPM in the
layman’s language.

2.0 Logic behind the CAPM

An asset is basically any instrument in the world which can be assigned a price and
can be traded. CAPM defines the market as the universe of all assets in the world.
CAPM is based on the portfolio theory which was the work of Harry Markowitz. This
theory models the uncertainty in the price of an asset by the variance of the returns
of the asset [σ2] where σ is the standard deviation of the returns of the asset. The
portfolio theory is the mathematical proof of the common sense adage “Do not put
all your eggs in the same basket” meaning reduce your risk by diversifying. For
those who are mathematically oriented, the Appendix I gives a mathematical proof of
the portfolio theory.

An investor faces two risks when he invests in assets. One is the asset specific risk.
For example this is the risk which one faces when one holds a Microsoft stock. The
other risk which the investor faces is the market risk. This is the risk which is
common to all securities. Say for example a recession might be a type of market risk
in which all stocks might get a beating. Market risk is also called as systemic risk.

From the portfolio theory we see that if we diversify enough, the only risk that we
face is the market risk and not the asset specific risk. In other words, what CAPM
suggests is that asset specific risk is diversifiable and can be eliminated and hence
need not be compensated for. So the only risk which needs to be compensated for is
the non diversifiable risk which is the market risk. So for example if we assume the
market to be S&P 500 stocks, the investor will not be compensated for holding each
stock but should be compensated for the risk he accepts by holding the whole S&P
500 portfolio.

3.0 The CAPM formula

To make the above ideas more concrete mathematically, we proceed as follows:

Dec 2006 Page 2 of 6 Hrishikesh Herekar

Let Rf be the risk free rate. Typically assumed to be the interest rate of US
government bonds of the same tenure as the time frame for which the CAPM is being
used.

E(Rm) is the expected rate of return on the market. Though CAPM defines the
market as the universe of all assets in the world, practically we consider some finite
proxy for the market. Say for example S&P 500 can be considered to be a good
proxy for the market for the US stocks. E(Rm) is typically calculated as historical
market returns for the last five to ten years.

The market premium is the expected market return over and above the risk free rate
viz. E(Rm) – Rf.

The CAPM states that, if an investor holds a market portfolio, then the risk
compensation for each asset should be commensurate with how that asset behaves
in relation to the whole market and not how it behaves individually. Thus for
example, if a stock co-varies exactly like the market, then its expected return should
be the same as the expected return from the market.

If Vm is the variance of the market (and is a proxy for market risk), then the risk
premium per unit of risk will be given as [E(Rm) – Rf] / Vm

So to find out the compensation for each asset, we need to see how risky each asset
is in relation to the market. In other words we need to see how each asset co-varies
with the market.

If E(Ri) is the expected return from an Asset i, then it should be equal to risk
premium per unit of risk multiplied by the contribution of the asset to the market risk
which is nothing but the covariance of the asset to the market.

E(Ri) = Cov(Ri,Rm) * [E(Rm) – Rf] / Vm

Even if the covariance of the asset with the market is zero, i.e. the asset is
essentially risk free, investors will still expect a return for the time that they stay
invested which is nothing but the risk free rate Rf.

Thus the complete formula for CAPM will be:

E(Ri) = Rf + Cov(Ri,Rm) * [E(Rm) – Rf] / Vm

Re-arranging we get,

E(Ri) = Rf + [Cov(Ri,Rm)/Vm] * [E(Rm) – Rf]

The factor [Cov(Ri,Rm)/Vm] is called the beta of the asset (β).

The resulting model is the CAPM and is typically written as:

E(Ri) = Rf + βi * [E(Rm) – Rf]

Dec 2006 Page 3 of 6 Hrishikesh Herekar

4.0 The Beta of an Asset

The beta of an asset is typically arrived at by regressing historical asset returns

against historical market returns. The slope of the line arrived at in the regression
gives the beta of the asset. It is fairly easy to calculate the beta of a stock as the
historical returns are easily available. The betas of stocks are normally published by
exchanges and other institutions such as yahoo finance which give stock information.

Fig 1: Calculation of beta for an asset

The beta is an indicator of how much a particular asset contributes to the risk of the
whole market. Obviously the beta of the market will be 1. If an asset is as risky as
the market, its beta will be 1. If it is riskier than the market its beta will be more
than 1. If it is less risky its beta will be less than 1. Risk free assets will have a beta
of 0.

5.0 Conclusion

The article tries to cover the basics of CAPM in layman’s terms. The CAPM is a very
vast subject and a lot has been written about it including its benefits and limitations.
The reader is requested to refer to a standard book after going through this article.
Also a lot of information is available on the net. A few websites are included in the
next section for further reading.

6.0 Material for Further Reading

1. http://en.wikipedia.org/wiki/Capital_asset_pricing_model
2. http://www.investopedia.com/terms/c/capm.asp
3. http://www.moneychimp.com/articles/risk/classes.htm
4. Principals of Corporate Finance by Brealey and Myers

Dec 2006 Page 4 of 6 Hrishikesh Herekar

7.0 Appendix I – The Portfolio Theory Derivation

Though this section is mathematical, it should not be difficult even for those with a
brief background in statistics. The section derives the result used in section 2.0.

Let us assume that there are only two assets in the world with variances V1 and V2
each having a weight of W1 and W2. The weights indicate how many units of the
asset makeup the portfolio. For example if the two assets are assumed to have equal
units in the portfolio, then W1 = W2 = 0.5.

Then the total variance in the market portfolio (Vm) will be:

Vm = W12 * V1 + W22 * V2 + 2 * W1 * W2 * Cov(R1, R2)

This comes from the simple statistical rule of addition of two variances.

Cov(R1,R2) is the covariance between the returns of the assets. For example if one
asset returns increase when the second asset returns increase then the covariance is
positive. But if one asset returns decreases when the second asset returns increase
then the covariance is negative. The covariance has a formula for calculation but we
need not bother about it at this stage.

Thus we see that when there are two assets, there are two individual variance terms
and one covariance term. We can generalize this for a market having “n” number of
securities. Then we will have “n” variance terms and nC2 co-variance terms.

Then the market variance Vm can be found out as:

Vm = W12 * V1 + W22 * V2 + … + Wn2 * Vn + 2 * W1 * W2 * Cov(R1, R2) + … + 2 *

Wn-1 * Wn * Cov(Rn-1, Rn)

To simplify things let us assume the following:

1. W1 = W2 = … = Wn = 1/n [That is each asset is equally weighted in the

portfolio]
2. V1 = V2 = … = Vn = V [That is all assets have the same variance]
3. All covariance terms are equal and are symbolically represented as Vmm.

Applying the above assumptions we get,

Vm = (1/n) 2 * V + … (n terms) + 2 * (1/n) * (1/n) * Vmm + ... (nC2 terms)

Thus Vm = (1/n) 2 * V * n + 2 * nC2 * (1/n) * (1/n) * Vmm

Canceling common terms & rearranging we get,

Vm = (1/n) * V + (1-1/n) * Vmm

Now as the number of assets in the market tends to infinity, 1/n will tend to zero.
Thus for infinite number of assets, Vm = Vmm.

Dec 2006 Page 5 of 6 Hrishikesh Herekar

This result proves that as we keep on adding assets to our portfolio, the risk of the
portfolio is independent of individual asset risks but is a function of the co-variances
between the assets. The asset specific risk goes on reducing as the individual risks
cancel out each other. For example if one stock is falling in price, some other stock
might increase in price and thus the individual risks cancel out. Theoretically if we
hold all the assets in the entire market, the risk that we are exposed to will be the
market risk and not the asset specific risk.

This can be represented diagrammatically as follows:

Risk

Risk of the portfolio

Market Risk

Number of assets in the portfolio

Fig 2: The portfolio theory – Risk of a portfolio goes down with diversification

Dec 2006 Page 6 of 6 Hrishikesh Herekar